Condoms - health economics report
A model to evaluate the cost effectiveness of
condom distribution (CD) schemes, developed
for NICE public health guidance on condom
distribution schemes and sexually transmitted
infections (STIs)
May 2016
Authors:
Susi Sadler
Jonathan Tosh
Andrew Rawdin
Hazel Squires
Anthea Sutton
Jim Chilcott
School of Health and Related Research (ScHARR)
University of Sheffield
Regent Court
30 Regent Street
Sheffield
S1 4DA
UK
Acknowledgements:
The ScHARR team would like to thank the following for their help with the work: Owen BrigstockBarron from PHE, Hannah Sowerbutts from Leeds City Council, Kathryn Faulkner from Norfolk And
Norwich University Hospitals NHS Foundation Trust and Andrew Evans from METRO for information
about the operation of a number of C-card schemes and information to help cost these schemes.
METRO also provided data to support this which was extremely useful in validating the costs used in
the model. Thanks go to Ford Hickson (Sigma Research, London School of Hygiene & Tropical
Medicine) for the latest condom use data from the 2010 and 2014 European MSM Internet Survey.
Also Amanda Mason-Jones from York University for feedback on the draft report and recommended
additional sources. Finally to Chris Carmona, Sarah Richards, Rebekah Pennington, Ian Wall and
Adrienne Cullum from NICE for on-going support and advice on the work.
2
TABLE OF CONTENTS
EXECUTIVE SUMMARY ............................................................................................................................ 6
1
2
BACKGROUND ............................................................................................................................... 10
1.1
Purpose of this report ........................................................................................................... 10
1.2
The role of economic evaluation within the NICE process ................................................... 10
1.3
Existing cost effectiveness evidence of condom distribution (CD) schemes ........................ 13
METHODS ...................................................................................................................................... 14
2.1
Overview of methods............................................................................................................ 14
2.2
Generic Model....................................................................................................................... 15
2.2.1
Economic model scope ................................................................................................. 15
2.2.2
General modelling methodology .................................................................................. 17
2.2.3
Model schematic ........................................................................................................... 18
2.2.4
Uncertainty ................................................................................................................... 19
2.2.5
General model parameters ........................................................................................... 20
2.2.6
Modelling quality adjusted life years (QALYs) .............................................................. 20
2.2.7
Modelling the cost of an STI averted ............................................................................ 21
2.2.8
Effectiveness evidence .................................................................................................. 24
2.3
2.3.1
Economic model scope ................................................................................................. 25
2.3.2
Effectiveness evidence .................................................................................................. 25
2.3.3
Model-specific parameters ........................................................................................... 26
2.3.4
Modelling sexually transmitted infections (STIs) .......................................................... 28
2.3.5
Exploratory analysis around pregnancy outcomes: young people ............................... 30
2.3.6
Costing the C-card scheme ........................................................................................... 31
2.4
3
Young People Model ............................................................................................................. 25
Men who have sex with men (MSM) model ......................................................................... 32
2.4.1
Economic model scope ................................................................................................. 32
2.4.2
Effectiveness ................................................................................................................. 32
2.4.3
2.5
Economic model scope ................................................................................................. 36
2.5.2
Effectiveness evidence .................................................................................................. 36
2.5.3
Model-specific parameters ........................................................................................... 36
2.5.4
Modelling sexually transmitted infections (STIs) .......................................................... 36
5
General population model .................................................................................................... 39
2.6.1
Economic model scope ................................................................................................. 39
2.6.2
Effectiveness evidence .................................................................................................. 39
2.6.3
Model-specific parameters ........................................................................................... 39
2.6.4
Modelling sexually transmitted infections (STIs) .......................................................... 40
2.7
4
Black African model .............................................................................................................. 36
2.5.1
2.6
3
Model-specific parameters ........................................................................................... 32
Summary of parameters ....................................................................................................... 42
RESULTS......................................................................................................................................... 44
3.1
Young People (C-Card) scheme ............................................................................................. 44
3.2
MSM scheme ........................................................................................................................ 54
3.3
Black African scheme ............................................................................................................ 58
3.4
General population scheme .................................................................................................. 62
DISCUSSION AND CONCLUSION .................................................................................................... 66
4.1
Discussion of results.............................................................................................................. 66
4.2
Limitations and further research .......................................................................................... 67
REFERENCES .................................................................................................................................. 69
FIGURES
Figure 1: The cost effectiveness plane .................................................................................................. 12
Figure 2: Conceptual model of the impact of CD schemes ................................................................... 14
Figure 3: Cost-effectiveness plane for base-case analysis of a multi-component (C-card) intervention
in young people (13-24 years), effectiveness per Larsson et al. (20). .................................................. 47
4
Figure 4: Cost-effectiveness plane for scenario analysis of a multi-component (C-card) intervention in
young people (13-18 years), effectiveness per Larsson et al. (20). ...................................................... 49
BOXES
Box 1: Incremental Cost Effectiveness Ratio ........................................................................................ 11
5
EXECUTIVE SUMMARY
Background
This report examines the potential economics of condom distribution (CD) schemes for the
prevention of sexually transmitted infections. It should be read in conjunction with a companion
report of the evidence for the effectiveness of such schemes prepared by The National Institute for
Health and Care Excellence (NICE).
Methods
An economic model is used to bring together data from a number of different sources in order to
estimate outcomes including sexually transmitted infections (STIs) avoided, quality adjusted life
years (QALYs) gained, costs for a total England catchment population, overall cost per QALY gained
and where relevant numbers of unintended pregnancies avoided as a secondary outcome. Cost and
health impacts for a lifetime horizon are estimated for different age groups and genders. Costs are
calculated from a public sector perspective. The analysis uses a simple steady state Bernoulli Process
model to estimate the impact of changes in condom usage on STI cases, which has been used in
previous NICE assessments including guidance on young people’s contraceptive services.
Probabilistic sensitivity analysis, scenario analyses and threshold analyses are used to explore the
economic outcomes.
The analysis focuses on CD schemes in four populations:
Multi-component CD scheme in young people, based around C-Card services:
A scheme targeted at young people aged 13-24 aiming at increasing condom use and reducing
condom failure. A base-case scenario population of young people aged 13-24 was assessed
using the effectiveness evidence from a controlled study in Sweden of a multi-component
intervention in a high-school setting (1) which found a relative risk of ‘ever using’ a condom of
1.23 in the intervention group compared to control. C-card scheme costs are estimated from
four published reports which give the costs of C-card schemes in England (2-5), and were within
a small range, giving a mean cost of £0.48 per person in the target intervention. In addition
three scenarios were tested;
1. With a narrower age group from 13-18 as the population (in line with the effectiveness
study population)
6
2.
With condom breakage reduced in the intervention group using a relative risk of 0.8
following a study by Macaluso et al.
(6), which showed breakage reducing with
increased experience in condom use
3. With higher prevalence of HIV (average UK prevalence rates)
A threshold analysis was carried out, to identify the combination of costs and effectiveness
required to make a scheme cost effective at £30,000 per QALY, or to make a scheme dominant
(i.e. QALY improving and cost-saving). An existing model developed for NICE was used to
explore the potential impact of this intervention upon pregnancy outcomes.
CD schemes for men who have sex with men (MSM), black Africans, and the general population
As there is limited effectiveness data for these 3 populations, threshold analysis was conducted
to identify the combination of costs and effectiveness data required for a CD scheme to be cost
effective or cost saving. This was conducted for each population (using population-specific
inputs), for low, central and high estimates of HIV prevalence.
Results
Condom distribution schemes for young people
The baseline analysis targeted at the 13-24 years is most representative of the existing C-card
scheme in the UK. The quality of life effects are heavily influenced by the impact of a small number
of HIV cases prevented. The model predicts that a scheme with costs in the region of a typical C-Card
scheme could be expected to have a cost per QALY gained in the region of £17,411 per QALY gained.
However, it should be cautioned that the evidence for effectiveness in this broader age group is not
demonstrated.
When C-Card is considered for the narrower age-group (13-18 years), which is most coherent with
the study population of the Larsson study (16-20 with mean age of 17), a mean cost per QALY of
£45,856 compared with no C-card is estimated, meaning it would not be cost-effective at the £2030,000 per QALY gained level. This is due to lower rates of sexual activity and underlying prevalence
of STIs in the younger age group.
Whilst reduced condom breakage might improve cost effectiveness, the impact is estimated to be
small (the cost per QALY gained reduces to around £14,469). In the scenario with higher prevalence
7
of HIV, the increase in HIV cases averted by CD schemes makes them cost-saving overall (£10m
savings compared with £3.5m scheme costs across England in the target population).
The analysis exploring the potential impact of the C-card on pregnancies suggests that the costs
saved from preventing or delaying a small number of pregnancies in those aged under 18 are
significantly greater than savings from STI prevention. For example, at a relative risk of 1.22 savings
from STIs are estimated to be less than £764,000 (compared with a scheme cost of £1,543,000) in
the cohort of English young people aged 13-18, whereas pregnancy-related savings are estimated at
around £11m (including government Benefits). If these pregnancy-related savings were to be
included within the cost effectiveness estimates (and excluding the impacts of pregnancy upon
QALYs), the base case scenario becomes cost-saving (total savings over £10m). If government
Benefits are excluded, the additional £318,000 savings shift the cost per QALY gained for the base
case to around £12,000.
Condom distribution schemes for men who have sex with men
If a scheme is effective on this sub-group the increased QALYs and savings from preventing HIV are
much larger than the impacts from other STIs. The threshold analysis suggests that even schemes
with relatively high costs per person can be cost effective if a small improvement in condom use can
be achieved and demonstrated. For example, even at the lowest HIV prevalence estimates, a scheme
which gave a relative risk of condom use of 1.04 compared with baseline would be cost effective
even at £10 per person in the target population (twenty times the cost suggested for a multicomponent C-card scheme).
Condom distribution schemes for black Africans
If a scheme is effective on this sub-group the increase in QALYs are mainly from HIV prevention,
although in terms of cost savings, HIV, PID and chlamydia prevention are all important. The
threshold analysis suggests that due to the higher prevalence of HIV in this sub-group, effective
schemes could be cost-saving at relatively high cost. For example, even at the lowest HIV prevalence
level, a scheme with relative risk of condom use of 1.04 would be cost effective even at £10 per
person.
Condom distribution schemes for the General population scheme
If a scheme is effective in this broad population the increase in QALYs are mainly from HIV
prevention, although in terms of cost savings, HIV, PID and chlamydia prevention are all important.
8
The threshold analysis suggests that schemes need to balance cost and effectiveness to be cost
effective. For example, at central HIV prevalence estimates, a scheme with relative risk of condom
use of 1.1 would have to cost around £1 per person in the target population or less to be cost
effective.
Limitations and further research
The economic modelling is subject to a number of limitations. The economic impact of CD schemes
was based on very limited effectiveness evidence, with a high degree of variability in CD scheme
designs, costs and effects being suggested by the available evidence. The model structure and
underlying assumptions may also have simplified the transmission of STIs.
Obtaining robust evidence is difficult because of the sensitive nature of the topic and the challenges
around conducting research in this area. Nonetheless, In order to understand the economics of CD
schemes in the UK it is imperative to have evaluations that demonstrate the impact of these
schemes on behaviour change and condom usage in UK-relevant populations.
9
1
BACKGROUND
1.1
Purpose of this report
The purpose of this report is to evaluate the cost effectiveness of a range of condom distribution
(CD) schemes to encourage the effective use of condoms, and subsequently reduce the cases of
sexually transmitted infections (STIs) in England. It accompanies the evidence review report on the
effectiveness of CD schemes undertaken by the team at the National Institute for Health and Care
Excellence (NICE).
This report considers the cost effectiveness of CD schemes to which reduce the cases and the direct
consequences of STIs. The report additionally includes an exploratory analysis of the impact that CD
schemes may have on unintended pregnancies, as the schemes evaluated may have this benefit,
either directly as a target of a CD scheme, or indirectly as a consequence of increased effective use
of condoms.
1.2
The role of economic evaluation within the NICE process
NICE was established in the UK in 1999 with a mandate from the Department of Health to appraise
the health benefits and costs of new and established health technologies and clinical practice. NICE
guidance now covers England, with the devolved UK nations having their own separate decision
making systems. In 2005, NICE’s remit expanded to include public health. When money is spent on
new health interventions within the health or wider public sectors, either existing interventions will
be displaced or the budget will expand. Therefore a rational and coherent framework is required to
help inform decisions about which interventions should be recommended for reimbursement.
An economic evaluation is a comparative analysis of the costs and consequences of two or more
competing interventions. There are several methods of economic evaluation specific to health. A
cost effectiveness analysis (CEA) values consequences in a natural unit of health (for example, life
years gained or case averted), a cost-utility analysis is a form of CEA which values consequences
using a generic measure of health (often the Quality Adjusted Life Year – QALY). A costconsequences analysis will compare a range of alternative outcomes against the net costs between
competing choices, but will not necessarily infer the value of these consequences to society. For any
economic evaluation, it is important that all consequences are captured, even when they extend far
into the future. An efficient decision is one where the benefits gained by the new intervention are
greater than the benefits forgone by services and interventions displaced due to any additional costs
being imposed on the constrained system.
10
An economic model is a method of objectively combining a range of data and evidence to inform an
economic evaluation of the cost and consequences of alternative interventions. Models also allow
extrapolation into the future, where required, and can quantify the uncertainty in the structure of
the model, and the parameters within.
In order to assess the impact of assumptions and uncertain evidence in a model and upon its results,
sensitivity analysis is a key stage of any economic evaluation. This involves varying assumptions and
parameters to assess their impact on the models results. If the results change significantly, then it
may be worth investing in generating more reliable evidence to inform the decision. Within a model,
all key uncertainties and assumptions should be tested to evaluate their impact on the model’s
results.
Because a CEA does not provide a direct comparison of the value of the effects and the costs,
decision rules are required. If comparing two treatments, one may cost more but also provide more
QALYs. The problem for decision-makers is determining how much extra benefit is required to justify
the extra expenditure, because there is an opportunity cost associated with allocating resources. A
cost effective treatment is one where, given limited resources, its use will contribute to the
maximisation of health benefits. Traditionally, the output of a CEA is reported as a ratio of the
difference in costs and difference in effects between two alternatives, the incremental cost
effectiveness ratio (ICER), as demonstrated in Box 1.
𝐼𝐶𝐸𝑅 =
(𝐶𝑎 − 𝐶𝑏 )
(𝐸𝑎 − 𝐸𝑏 )
[1]
∆𝐶
∆𝐸
[2]
Where:
𝑎 = Intervention A
𝑏 = Intervention B
𝐶 = Costs
𝐸 = Effects
𝐼𝐶𝐸𝑅 =
Box 1: Incremental Cost Effectiveness Ratio
The ICER can be interpreted as the incremental cost per incremental unit of effect, as represented in
equation [1]. For decision makers, there are six possible situations when using an ICER presented in
Figure 1.
11
Incremental
cost
£150,000
Cost effectiveness
threshold
£100,000
D
C
£50,000
B
QALYs gained
£0
-6
-4
-2
E
0
2
4
6
-£50,000
-£100,000
F
A
-£150,000
A
New intervention dominates conventional therapy (i.e. more effective and cost saving)
B
New intervention more effective but greater cost (cost effectiveness better than
threshold) compared to conventional
C
New intervention more effective but greater cost (cost effectiveness worse than
threshold) compared to conventional
D
New intervention dominated by conventional (ie less effective and greater cost )
E
New intervention less effective but cost saving (cost effectiveness worse than threshold)
compared to conventional
F
New intervention less effective but cost saving, (cost effectiveness better than threshold)
compared to conventional
Figure 1: The cost effectiveness plane
With situations A and D, the solution for decision-makers is very straight forward. With situation A,
the new intervention is better than the conventional therapy and is also cost saving, the new
intervention is said to dominate the conventional therapy and clearly provides good value for
money. In B the new intervention is less effective and costs more than the conventional therapy, is
said to be dominated by the existing therapy and is clearly not good value for money.
In situations B and C, the new intervention is better than the existing therapy but costs more. In
situation B the incremental cost per unit benefit gained (in most cases the QALY gained) is better
than a notional threshold value, and in situation C the cost effectiveness in worse than the threshold
In situations E and F the new intervention is less effective than the conventional therapy but cost
saving. In situation F the cost saved per unit of health benefit foregone implies that the cost
effectiveness might be considered favourable, that is reinvesting this cost saving elsewhere may
realise greater health benefits elsewhere in the system. By contrast in situation E the cost
12
effectiveness is worse than the conventional therapy. The cost effectiveness evidence is one of
several factors that determine a positive or negative decision by NICE.
Within the current analysis, health outcomes are considered and valued using a QALY. Also
presented are absolute numbers of cases of STIs averted due to an effective condom distribution
(CD) scheme.
1.3
Existing cost effectiveness evidence of condom distribution (CD) schemes
NICE carried out a review of the effectiveness evidence of CD schemes. The full review is available
elsewhere (7). The review identified 20 studies across a range of settings (healthcare, schools,
outreach and community) and in both single-component (condom availability only) and multicomponent (availability with education, advice or support) styles. Overall, the quality of the studies
was considered to be poor, with only one study meeting all or most of the checklist quality
assessment criteria and 6 meeting some of the criteria. In addition, 16 of the studies were US-based
and only two were based in the UK.
Overall there was some limited evidence from three studies showing that multi-component schemes
in high schools can increase condom use at last intercourse without increasing sexual activity. There
was also some evidence that multi-component community or outreach schemes targeted at highrisk individuals can increase condom use. There was mixed evidence around the effectiveness of
multi-component schemes in healthcare settings, with some evidence that schemes increased
condom acquisition but no evidence of increased usage.
There was mixed evidence of the
effectiveness of single-component schemes in high schools, with some studies reporting increases in
condom use and others reductions or no change. There was no evidence of the effectiveness of
single component schemes targeted at high-risk individuals in increasing condom use.
13
2
METHODS
2.1
Overview of methods
A conceptualisation of the decision problem was undertaken to help understand the potential
impact of CD schemes on population health and services, including NHS, public and third sector
providers.
A conceptual model was developed based on an initial mapping review to inform what might be
included with the economic model. The conceptual model is provided in Figure 2.
Impact on
screening rates
(e.g. National
Chlamydia
Screening
Programme
(NCSP))
Impact on ‘unplanned’
pregnancies
CONDOM
Enhanced risk
perceptions, and
normalisation of
condom use
DISTRIBUTION
SCHEME
Impact on other contraceptive
services
Impact on peers/social
network
STI rate
Identified
infection
Unidentified
infection
Long term
health
Long term
health
Treatment
Costs
Legend
= Service impact
= Health impact
= Intermediate impact
Figure 2: Conceptual model of the impact of CD schemes
14
Transmission of
infection
Based on this conceptual model, and the review of effectiveness and economic evidence, a health
economic model has been developed. From this model, CD schemes are applied to provide an
incremental comparison of the costs and consequences of alternative CD schemes, compared to no
CD scheme. This model, ‘the general model’, is explained in the next section.
2.2
Generic Model
2.2.1 Economic model scope
2.2.1.1
Population
A general model was created, from which models for specific sub-groups could be developed. The
primary hypothetical population of the general model is all people at risk of a sexually transmitted
infection (STI). This is deemed to be all people in England who are sexually active. The general model
has a national (England) population ranging from 13-90 years of age. The age bounds of the
population can be modified so that the costs and consequences of a scheme that targets a particular
sub-population can be clearly identified and quantified. These sub-groups include people who are at
the greatest risk of acquiring an STI, are defined within the model by their age, gender and
background prevalence of STIs. This population can be further defined by changing model
parameters controlling their number of sexual contacts, and condom usage, along with background
prevalence of particular STIs. This enables evaluations of CD schemes in subpopulations of particular
interest. We report results of analyses for four specific groups of interest:
1. Young people (ages 13-24 in the base case)
2. Men who have sex with men (MSM)
3. Black Africans
4. The general population
Prisoners are excluded from the scope due to being covered by the NICE guideline on the physical
health of people in prisons that is currently being developed.
The use of condoms for contraception will not be considered in the base case analysis. It is however
recognised that CD schemes may also directly or indirectly reduce unintended pregnancies, and so
an analysis is conducted exploring the impact of CD schemes on unintended pregnancies in young
people. The economic model does not capture additional benefits of reducing vertical transmission
from mother-to-child of HIV or syphilis, where severe consequences for newborns (congenital
syphilis, paediatric HIV infections) could be averted in both of cases. Furthermore, syphilis in
pregnancy is associated with foetal death, spontaneous abortion, stillbirth or premature birth, low
15
birth weight due to intrauterine growth restriction and perinatal death, as well as serious sequelae
in live born infected children. Consideration of these benefits may result in additional cost savings
and QALY gains.
2.2.1.2
Intervention and comparators
The model was used to assess four population groups,:
-
Young people: to understand the impact of C-card; a multi-component scheme which
provides free condoms, with or without lubricant, together with training, information and
other support.
-
Men who have sex with men: to understand the impact of CD schemes that provide or
distribute free condoms and lubricant, specifically where these target MSM.
-
Black Africans: to understand the level of effectiveness required to make a scheme cost
effective in this sub-group with higher HIV prevalence.
-
General population: to understand the impact of CD schemes available to the general public,
providing free or cheap condoms and lubricant.
Details of the general model are set out in this section. Specific details of the separate modelling
methods for each of these sub-groups are set out in sections 2.3, 2.4, 2.5 and 2.6 respectively.
Within the economic model, it is assumed that a CD scheme will have a per capita cost for the
population or sub-population of interest, and will result in an absolute increase in the proportion of
the population who regularly use condoms. This will be compared to the current background use of
condoms without a particular CD scheme.
2.2.1.3
Outcomes
Outcomes are presented in terms of the number of specific STIs averted within a population. The
STIs modelled are:

Chlamydia

Gonorrhoea

Human Immunodeficiency Virus (HIV)

Syphilis
Also modelled is Pelvic Inflammatory Disease (PID), an infection that occurs in some women who
have either chlamydia or gonorrhoea.
16
The following outcomes will be presented:

Cost per Quality Adjusted Life Year (QALY) gained – Each STI case averted will correspond to
a cost and QALY gain. For some STIs, this gain may be very small; however for others, such as
HIV, the gain is likely to be significant.
The cost of an STI case assumes that a person with a symptomatic infection will present at either a
GP surgery or a Sexual Health or Genitourinary Medicine (GUM) clinic. The cost includes this
presentation, any tests required, treatment and any follow-up consultations required.
2.2.1.4
Perspective
CD schemes may be provided by the public, private or charitable sector. A public sector perspective
will be taken for the economic analysis. This is in line with the NICE methods for guidance
development (8).
2.2.1.5
Discounting
The analysis uses a lifetime horizon for costs and QALYs for cases of HIV averted. Costs are
discounted at 3.5% per annum, QALYs are discounted at 3.0% per annum due to the data available
for this parameter.
2.2.2 General modelling methodology
The model was developed within Microsoft Excel. The model enables the economic evaluation of a
CD scheme by comparing its net costs incurred and benefits accrued relative to current practice.
A per-capita per-year cost of a CD scheme is introduced to the model, and a subsequent increase in
the proportion of people using condoms will result in a decrease in the number of cases of STIs in
one year. The model does not have a time-varying/dynamic process but rather estimates directly a
‘steady state’ STI acquisition rate under the status quo and CD scheme scenarios and associated
cost-effectiveness, which also enables the budget impact of any CD scheme to be clearly
demonstrated for implementation purposes.
The model uses an established Bernoulli Process model (now called the “STI model” for clarity),
which estimates the number of STIs in a cohort of people (9). The STI model has been used in other
economic evaluations in related areas (10), including the NICE Public Health guideline for
Contraceptives (11), and the discontinued Sex and Relationship Education guideline(12)
17
The STI model assumes that the proportion of the population who acquire a particular STI is a
function of the prevalence of an STI, the proportion of the population who use condoms, the
transmission rate of the STI, the condom failure rate, and the number of sexual contacts (per year).
𝑊 = 𝑣((𝑔(1 − ((1 − 𝑡𝑘)𝑠 ))) + ((1 − 𝑔)(1 − ((1 − 𝑡)𝑠 ))))
Where:
𝑊= Proportion of population acquiring an STI
𝑣 = Prevalence of an STI
𝑔 = Proportion using condoms
𝑡 = Transmission rate
𝑘 = Condom failure rate
𝑠 = Acts of sexual intercourse per annum
Using this STI model allows the proportion of people using condoms (𝑔) to be adjusted, while all
other parameters are held constant (for the specific population under consideration). Increasing 𝑔
will reduce the number of STI cases acquired, which enables a CD scheme to be evaluated via this STI
model estimating the total number of STI cases averted and sensitivity analyses exploring the impact
of the different scheme in population subgroups etc.
The sources of data for each of these parameters are reported in the subsequent sections.
The model is not a dynamic infection transmission model, due to data and resource constraints. The
appropriateness of this model and the simplifications that it requires should be considered alongside
the models results. These are discussed further in the discussion section.
2.2.3 Model schematic
A model schematic is provided in Figure 3 which shows how costs and outcomes are estimated in
the model. A whole population of people is modelled for one year. Proportions of this population
will develop an STI, which will incur health care costs and dis-benefit health (QALYs). For the CD
scheme arm of the model, there is a per-patient-per-year cost (an incremental cost compared to
current practice) for providing the scheme to the population. The effectiveness of the CD scheme is
18
modelled as an absolute increase in condom use, and the subsequent reduction in the number of STI
cases.
STI cases
Costs
STIs
QALYs
Net costs
CD Scheme
Incremental cost of
programme above
current practice
Non-cases
Population
STI cases
Costs
STIs
QALYs
Net costs
Current practice
Non-cases
Figure 3: Decision model schematic
2.2.4 Uncertainty
A probabilistic sensitivity analysis (PSA) was conducted to assess the impact of parameter
uncertainty upon the model output. This involves assigning an appropriate statistical distribution to
uncertain parameters in the decision model. A random draw from each distribution is taken
simultaneously, and the model run a large number of times (5,000 was found to reach reasonable
stability whilst being feasible in terms of run time) to quantify the parameter uncertainty in the
model. This quantification can be represented as the probability of an intervention being costeffective at a particular ICER threshold, λ, and visually using cost effectiveness acceptability curves
(13). All results presented, base cases, scenarios and threshold analyses, are probabilistic.
19
2.2.5 General model parameters
2.2.5.1
Sexual activity and contacts
Age and gender-specific data for the proportion of people who are sexually active, and the
corresponding rate of sexual contacts for people over 16 were taken from the third National Survey
of Sexual Attitudes and Lifestyles (NATSAL-3) publication (14). These data are provided in Table 1.
For those between 13 and 16, the NATSAL-3 dataset was analysed to provide these data. It was
assumed that those under 13 are not sexually active. Data were not available on the number of
sexual contacts (if sexually active) if under 16 years old, and so it was assumed that this would be
equal to those aged 16-24. These data are generic and apply across all three modelled groups.
Table 1 Sexual activity and contacts, by age and gender
Age
13
14
15
16-24
25-34
35-44
45-54
55-64
65+
Sexually active
Men
Women
Sexual contacts (if active) per 4 Source
weeks)
Men
Women
4.4%
11.8%
26.0%
75.9%
90.1%
92.5%
86.4%
76.3%
59.8%
2.3%
8.5%
21.4%
77.0%
91.8%
80.8%
85.0%
63.7%
42.1%
5.11
5.11
5.11
5.1
5.4
4.1
4.1
3.2
2.3
5.81
5.81
5.81
5.8
4.9
4.0
3.5
2.5
1.4
NATSAL-3 dataset
Mercer 2013 Lancet3
1
Assumed equal to 16-24 data
2.2.6 Modelling quality adjusted life years (QALYs)
Depending on the particular STI, QALYs were modelled in one of two ways. Either an absolute QALY
reduction due to an STI was found in the literature and applied in this model. Or a disutility for an STI
was found, and multiplied by an estimate of time from symptom onset to cure, to estimate the QALY
reduction. All values and sources are provided in Table 2. The data used for HIV from Farnham et al.
(2013) (15) assumes people are diagnosed when infected, with a CD4 count above 500, and is
discounted at 3% per annum due to being an American study (compared to 3.5% per annum which is
recommended by NICE).
20
Table 2 QALYs for STIs
STI
Disutility
(95% CI)
Chlamydia
Gonorrhoea
HIV
0.103
0.103
Syphilis
Primary 0.0072
18.7
(0.0065, 0.0079)
Secondary 0.041 (0.036,
0.045)
Primary:Secondary
–
70:30
0.1
13
PID
Condition
length
(weeks)
1
2
QALY gained
(95% CI)
Source
0.002
0.004
4.45
(4.34–4.47)
0.0026
(0.0023–0.0028)
Turner et al. (2013) (16)
Turner et al. (2013) (16)
Farnham et al. (2013)
(15)
Tuite et al. (2014) (17)
Suktankar (2014) (10)
0.025
Looker et al. (2015) (18)
2.2.7 Modelling the cost of an STI averted
All costs are in 2015 prices. For each STI, the cost of a ‘case averted’ was assumed to be equivalent
to self-identification of an STI and attendance at a GP surgery or GUM clinic, along with the cost of
any diagnostic test and subsequent treatment. For all STIs except HIV and PID, this cost is simply a
cost of treatment with no expected follow-up costs for further health intervention. For HIV, there is
a much more significant lifelong cost of treatment, health care consultations and follow-up costs. For
PID there is the risk of further complications, and so external published data have been used (12).
This assumption may be an underestimate of the true cost of a case being averted, due to the
significant health resources spent on promotion, screening and public health interventions to reduce
the total number of STIs, as well as the total number of undiagnosed STIs. However, it is not possible
to fully quantify all of these diagnostic, screening and public health interventions and attribute them
to each case of a particular STI, and therefore they are not included within this analysis.
2.2.7.1
Modelling the cost of an HIV case averted
In 2011 the Health protection agency published a report which stated that Investing in the
prevention of HIV infection should be a priority as it was estimated that the lifetime cost saving per
infection prevented was be between £280,000 and £360,000 (19).The same value taken from a 2009
report published by the same agency was used by the Faculty of Public Health of the Royal College of
physicians in their submission to the House of Lords select committee on HIV and AIDS in the UK
(20). Public Health England published similar reports in 2012, 2013, 2014 and 2015 but none of them
contained an estimate of the lifetime treatment cost associated with HIV infection.
21
In 2015 a paper published by Nakagawa et al (21) the lifetime cost of HIV treatment was estimated
to be £360,800 which when discounted at 3.5% per annum would fall to £185,200. However, the
authors suggested that switching to generic drugs once pharmaceutical patents expire could reduce
the lifetime cost of treating a 30 year old gay male infected with HIV in 2003 who lived till the age of
72 years by more than half to £179,600 which again once discounted at 3.5% per annum would fall
to £101,200. These costs are based on the assumptions that the standards of care are those
specified in the 2012 edition of the guideline published by the British HIV association and assuming
that generic drugs would only cost 20% of the branded versions. The cost year used in the analysis is
2013 (assuming this is listed in the Unit Costs of Health and Social Care 2015 (22) as the cost year
2012/13) then inflating these costs to 2014/15 result in costs of £368,084; £188,939; £183,226 and
£103,243.
In the analysis we use the cost of £103,243 and assume a 95% confidence interval of £82,594 to
£123,892.
2.2.7.2
Modelling the cost of a PID case averted-
For this analysis, the total cost of a case of PID was taken from a report to the National Collaborating
Centre for Women’s and Children’s health that considered the full cost of a case of PID, including
future complications and hospitalisations (12) . The mean total cost for PID was estimated to be
£2,846 assuming that the cost year used in the report is 2009 then (assuming this is listed in the Unit
Costs of Health and Social Care 2015 (22) as the cost year 2008/09) then inflating these costs to
2014/15 result in costs of £3,124 since no confidence interval is placed around these costs we have
assumed that the limits of the 95% confidence interval are +/-20% of the mean cost.
2.2.7.3
Appointment cost
For each of the remaining STIs, it is assumed that in each case a person will develop symptoms and
present at their GP. A sensitivity analysis is that they will present at a Sexual Health or Genitourinary
Medicine (GUM) clinic. The costs of an appointment at one of these services are provided in Table 3.
Table 3 Cost of an STI averted: Appointment costs
Cost component
Unit Cost (£)
Source
GP Appointment
67.90
Unit Costs of Health and Social Care 2015£
GU Medicine appointment 62.39
National Schedule of Reference Costs 2013/14$
£
11.7 min consultation, excluding admission costs
$
WF01B (Non-admitted Face to Face Attendance, first. Non-consultant led outpatient)
22
2.2.7.4
Testing cost
For some STIs, diagnosis can be confirmed by an examination; however other may require one or
more diagnostic test. The testing regimen and cost for each STI are provided in Table 4.
Table 4 Cost of an STI averted: Diagnostic testing
STI
Test
NAATs
£6.84£
X
X
Total cost
Cell Culture
£6.84£
Blood
£3.00$
Cytology
£7.77%
Chlamydia
£6.84
Gonorrhoea
X
£13.68
Syphilis
X
X
£9.84
£
NHS Costs of Diagnostic Services – DAPS07 - 2013/14 National Schedule of Reference Costs
$
NHS Costs of Diagnostic Services – DAPS05 - 2013/14 National Schedule of Reference Costs
%
NHS Costs of Diagnostic Services – DAPS01 - 2013/14 National Schedule of Reference Costs
2.2.7.5
Treatment cost
A micro-cost estimation of the treatment for each STI was conducted. Because the treatment is
often antibiotic therapy, there are a range of different regimens and antibiotics which may be used.
For each STI, the appropriate NHS regimen has been costed using BNF 2016 drug prices, and the
mean, minimum and maximum cost of treatment estimated. Some STIs have alternative treatment
regimens, and in these cases the cost has been averaged. The treatment costs are summarised in
Table 5.
Table 5 Cost of an STI averted: Treatment cost
Condition
Chlamydia
Gonorrhoea
Syphilis
2.2.7.6
Treatment cost (£, 2016)
Mean Cost Minimum Maximum
2.92
0.48
11.91
8.98
4.76
16.71
13.82
1.65
25.54
Total cost of an STI case averted
As well as the initial appointment, there may be subsequent follow-up appointments to either
initiate treatment after a positive diagnostic test result, or to ensure curation after treatment.
Minimum and maximum number of visits were determined based on NHS treatment pathways for
STIs, and averaged to provide the expected number of visits for each STI (see Table 6).
23
Table 6: Cost of an STI averted: Number of appointments
STI
Chlamydia
Gonorrhoea
Syphilis
Expected number of visits
1.5
2.5
2.5
Minimum
1
2
2
Maximum
2
3
3
The total cost per STI case averted, as well as upper and lower bounds for sensitivity analysis, are
provided in Table 7. The minimum and maximum estimates are based on the maximum and
minimum number of visits, treatment costs, and the maximum estimate includes GUM
appointments rather than GP practice appointments.
Table 7 Cost of an STI averted: Total cost
STI
Cost (£)
Mean
Minimum
Chlamydia
121.92
75.76
Gonorrhoea
206.17
129.24
Syphilis
210.59
133.66
£
Maximum includes GUM appointment cost
Maximum£
166.58
280.61
285.03
2.2.8 Effectiveness evidence
A systematic review of comparative assessments of condom distribution schemes has been
undertaken by NICE and is reported separately. The systematic review focuses on single and multicomponent condom distribution schemes in a range of settings including high school (secondary)
education, healthcare, commercial and community and outreach. The effectiveness evidence from
each setting and specifically how it relates to the potential economic impact of CD scheme in each
setting is discussed in the specific sections relating to each model below.
Details of the specific models, including their parameters are set out in the following sections
24
2.3
Young People Model
2.3.1 Economic model scope
This model was intended to capture the effects of implementing a C-Card scheme. These schemes
provide young people free access to condoms (and in some schemes lubricant) through a range of
outlets (for example, pharmacies, community centres, clinics and schools) using a personal card. The
precise age range for C-Card schemes varies by implementation. The base case analysis uses a target
age range of 13-24, however the effectiveness evidence is taken from a study with a slightly younger
population and thus a scenario analysis for ages 13-18 years is presented. In order to register for the
card the young person must attend a one-to-one session with a trained provider, enabling them to
talk through sexual health and relationship issues, receive a condom demonstration and discuss
contraception. For under 16’s this includes an assessment of Fraser competence. Unlimited supplies
can be accessed, but the young person must undergo further periodic one-to-one meetings to
maintain access. It is hypothesised that this scheme would not only increase condom usage, through
both availability and education, but also improve correct condom usage and therefore reduce
condom failure rates through breakage and slippage.
2.3.2 Effectiveness evidence
The review undertaken by NICE identified a number of studies of multi-component interventions for
young people. However, these were all limited to school-based studies and do not fully represent
the kind of offering that a C-CARD scheme provides. Evidence from these studies was somewhat
mixed as regards the effectiveness of multi-component schemes. A quasi-experimental study (23)
identified a small significant effect on increasing reported condom use at last intercourse (OR 1.36)
in school pupils in the USA using an intervention which combined education and free condom
availability.
A before and after study (24) reports a non-significant increase in the reported use of a condom at
last intercourse (OR 1.12) amongst school pupils in the USA using an intervention which involved
abstinence counselling, availability of sexual health advice and free condom availability.
A good quality controlled study (1) identified a larger, significant increase in reported ‘ever having
used’ a condom (OR 2.11) amongst school pupils in Sweden, using an intervention which involved a
number of lessons about contraception and condom use with free condom availability.
The highest quality study, that by Larsson et al. (1) was chosen to give the estimated level of
effectiveness in the base case scenario. A relative risk of condom use of 1.23 was applied to baseline
25
condom use in the intervention scenario to give the increased rate of condom use after the
intervention.
Because this effectiveness evidence was not specific to the C-card scheme, it is feasible that the
effectiveness of the C-Card itself is somewhat different. Although the C-card scheme has a set of
generic implementation principles, each implementation is intended to be designed locally with
participation of the target population, costs are, therefore, quite specific to each locality running the
scheme. In order to improve the understanding of the cost effectiveness of the C-card under a range
of possible cost and effectiveness scenarios, a threshold analysis was carried out, to identify the
combination of costs and effectiveness required to make a scheme cost effective at £30,000 per
QALY, or to make a scheme dominant (QALY improving and cost-saving ).
A scenario is included where condom failure due to breakage rate is reduced, with an odds ratio of
0.8, based on the results of Macaluso et al., (6) discussed below.
2.3.3 Model-specific parameters
2.3.3.1
Condom usage and failure
Age-specific data for the proportion of people routinely using condoms were taken from an Office of
National Statistics (ONS) publication. (25) It was assumed that under-16s were the same as 16-19
year olds. These data are provided in Table 8. All these studies date from prior to 2000, but despite
varying results, there is reasonable evidence that failure rates reduce with experience. None of the
measures closely matches the experience gained through C-card, which is based on one-to-one
training, whereas the studies report on experience gained over time of using condoms. Therefore we
would have to assume that the instruction received through C-card registration, and availability of
free condoms speeds up the process of familiarisation and practising with condoms in advance of
them being needed for sex with a partner. The most appropriate study was Macaluso et al. because
it compared usual use with new users (as opposed to other studies which compared those who had
been using condoms for several years), and included both breakage and slippage in its failure rates,
hence this study was chosen, giving, an overall failure rate of 3.62% (2.3% breakage and 1.3%
slippage). In the base case, no change in failure rates will be assumed. In a scenario, breakage rates
will be adjusted in accordance with the results of this study (Odds ratio 0.8).
26
Table 8 Condom use (percentage by age)
Age
16-19
20-24
25-29
30-34
35-39
40-44
45-49
Condom use (%)
54%
54%
41%
46%
27%
10%
13%
Source
Lader & Hopkins. ONS
Contraception and Sexual
health report, 2008/9 (25)
Several studies were identified which give differential breakage and slippage rates by experience.
Macaluso et al. (6) reported slightly lower odds of failure amongst those with usual past condom use
compared with those who rarely or never used a condom (OR = 0.8). Messiah et al. (26) reported
higher failure rates for people with fewer years of condom use (7.8% for less than 5 years use and
1.4% for over 5 years use). In the same study rates by age and by years of sexual experience did not
follow a similar pattern. Frezieres et al. (27) reported lower breakage and slippage rates (0.9%) with
latex condoms over 6 months of recorded use in a trial, than in the first 5 uses (1.6%). Vessey et al.
(28) reported higher failure rates per 100 woman-years in groups with longer experience using
condoms (6.0% if <24 months, 4.0% if 25-48 months and 3.6% if 49+ months).
All these studies date from prior to 2000, but despite varying results, there is reasonable evidence
that failure rates reduce with experience. None of the measures closely matches the experience
gained through C-card, which is based on one-to-one training, whereas the studies report on
experience gained over time of using condoms. Therefore we would have to assume that the
instruction received through C-card registration, and availability of free condoms speeds up the
process of familiarisation and practising with condoms in advance of them being needed for sex with
a partner. The most appropriate study was Macaluso et al. because it compared usual use with new
users (as opposed to other studies which compared those who had been using condoms for several
years), and included both breakage and slippage in its failure rates, hence this study was chosen,
giving, an overall failure rate of 3.62% (2.3% breakage and 1.3% slippage). In the base case, no
change in failure rates will be assumed. In a scenario, breakage rates will be adjusted in accordance
with the results of this study (Odds ratio 0.8).
27
2.3.4 Modelling sexually transmitted infections (STIs)
As well as data regarding the sexual activity and condom utilisation in a population, to determine the
number of STI cases in a population, the background population prevalence and transmission rate of
each STI are required.
2.3.4.1
Prevalence
For Chlamydia, Gonorrhoea and Syphilis the number of cases for 2014, stratified by age group and
gender, was taken from data published by Public Health England1. For HIV data was obtained from
the HIV New Diagnoses and Death Database2, with the gender split for each age assumed to be prorata with the overall sex ratio. This data is presented in Table 9.
It was assumed that the numbers given in the Public Health England data referred to the number of
cases reported in each age group and therefore the data was divided by the number of years in each
age group and further divided by the population in each at risk group to give the percentage
prevalence of Chlamydia, Gonorrhoea, HIV and Syphilis by age group and gender for young people.
The PSA is based upon beta distributions according to the case numbers presented in Table 9, giving
mean rates presented in Table 10.
The base case estimates for HIV prevalence are based on diagnosis rates and so may underestimate
the true prevalence of HIV. Therefore, a higher HIV prevalence scenario was also modelled, based on
the estimated overall prevalence of HIV in the UK (not available by age) of 0.19% stated in the 2015
HIV Situation Report (29). This was assumed to follow the same pattern by age and sex as HIV
incidence, giving the prevalences shown in Table 11.
1
Public Health England. Table 2: STI diagnoses & rates by gender, sexual risk and age group, 2010 –
2014.London. Public Health England, Colindale
2
Public Health England. United Kingdom National HIV surveillance data tables. London. Public Health England,
Colindale
28
Table 9: Prevalence of STI cases 2014 from Public Health England data
Chlamydia
Age
group
Male
56
14,724
32,921
25,869
7,279
3,806
253
13–14
15–19
20–24
25–34
35–44
45–64
65+
Female
817
41,995
47,768
23,299
4,322
1,576
45
Gonorrhoea
Male
7
1,693
6,031
10,738
5,077
2,783
156
Female
60
2,629
2,853
2,048
525
245
9
HIV
Male
24
545
1,532
1,192
841
319
157
Syphilis
Female
8
182
512
398
281
107
53
Male
0
61
396
1,319
1,245
963
45
Female
0
32
49
102
43
34
3
Total at risk population
Male
628,933
1,670,664
1,829,362
3,672,381
3,559,027
6,712,115
4,185,891
Female
598,876
1,584,088
1,774,376
3,694,976
3,600,040
6,885,167
5,119,288
Table 10: Mean prevalence values used in the model for Chlamydia, Gonorrhoea, HIV & Syphilis in the young people
analysis.
Chlamydia
Male
Female
0.000%
0.000%
0.009%
0.136%
0.881%
2.651%
1.800%
2.692%
0.704%
0.631%
0.205%
0.120%
0.057%
0.023%
0.006%
0.001%
Age
group
0 – 12
13 – 14
15 – 19
20 – 24
25 – 34
35 – 44
45 – 64
65+
Gonorrhoea
Male
Female
0.000%
0.000%
0.001%
0.010%
0.101%
0.166%
0.330%
0.161%
0.292%
0.055%
0.143%
0.015%
0.041%
0.004%
0.004%
0.000%
HIV
Male
0.000%
0.004%
0.016%
0.015%
0.042%
0.033%
0.017%
0.004%
Female
0.000%
0.001%
0.006%
0.005%
0.014%
0.011%
0.006%
0.001%
Syphilis
Male
Female
0.000%
0.000%
0.000%
0.000%
0.004%
0.002%
0.022%
0.003%
0.036%
0.003%
0.035%
0.001%
0.014%
0.000%
0.001%
0.000%
Table 11: High prevalence scenario (young people) – overall HIV prevalence = 0.19%
Age group
13 – 14
15 – 19
20 – 24
25 – 34
35 – 44
45 – 64
65+
2.3.4.2
HIV
Male
0.054%
0.229%
0.209%
0.586%
0.471%
0.243%
0.053%
Female
0.019%
0.081%
0.072%
0.195%
0.155%
0.079%
0.014%
Transmission rate in the absence of condom use
Estimating the probability of transmission between an infected person and an uninfected person
from one unprotected sexual contact is difficult due to the ethical implications of this research.
Therefore the data for this important parameter are in some cases are old and limited. Rates and
sources are presented in Table 12.
29
Table 12 Transmission rate
STI
HIV
Chlamydia
Gonorrhoea
Syphilis
2.3.4.3
Transmission rate (per 1 sexual contact)
Men
Women
0.12%
0.39%
45%
45%
53%
53%
61.8%
61.8%
Source
Boily et al. (2009) (30)
Wang et al. (2000) (31)
Wang et al. (2000) (31)
Alexander
et
al.
(1949)(32)
Pelvic Inflammatory Disease
Pelvic Inflammatory Disease (PID) is an infection that often occurs during a case of chlamydia or
gonorrhoea in women. Rather than estimate cases of PID independently it was considered as a
function of the number of cases of Chlamydia and the number of cases of Gonorrhoea. PID can have
a significant impact on a women’s quality of life and in some cases can be a serious condition
requiring hospitalisation and surgery. Data were found providing the proportion of chlamydia and
gonorrhoea cases that result in PID (see Table 13).
Table 13 STIs that lead to PID
STI (leading to PID)
Chlamydia
Gonorrhoea
Proportion
16% (6 – 25%)
0.9% (0.1 - 1.8%)
Source
Price et al. (2013) (33)
Unpublished conference abstract (Australian data)
(9)
2.3.5 Exploratory analysis around pregnancy outcomes: young people
In order to provide an indication of the relative impact of the interventions upon pregnancy
outcomes compared with STI outcomes, an exploratory analysis has been undertaken. This is limited
to interventions targeted at young people (14 – 18) because it makes the simplifying assumption
that all pregnancies within this population are unintended. It would not be possible given the data
available to distinguish between intended and unintended pregnancies, and therefore to assess the
potential impact of the interventions within the general population. This exploratory analysis uses an
existing model developed for NICE Public Health Guidance 511.
1
A detailed description of this model can be found at:
https://www.nice.org.uk/guidance/ph51/documents/contraceptive-services-for-socially-disadvantaged-youngpeople-additional-consultation-on-the-evidence-cost-effectiveness-modelling-report2
30
The analysis presented here assumes that the intervention targets people aged 14 – 18 years, and
that all pregnancies within this age group are unintended. It is assumed that there is a 50%
probability that the intervention delays the pregnancy and a 50% probability that the intervention
prevents the pregnancy. Within the PSA, this probability is varied from 0% to 100%. For those
pregnancies which are delayed, they are assumed to be delayed until the person is aged 19 – 24
years. The delayed births are equally divided over this age range.
Within the model, the probability of becoming pregnant by age and the probability of having an
abortion by age have been updated using the latest national statistics (34). The model has been
updated to use the same condom failure rate as the STI model, with a mean of 3.6% (6). The cost
savings predicted by the model have also been uplifted to 2014/15 prices (35).
2.3.6 Costing the C-card scheme
To cost the C-card scheme a rapid search identified a small number of published documents from
local schemes in England and Wales, of which 5 provided overall costs of their schemes: Derbyshire
(5), Lincolnshire (3), Nottinghamshire (2), Newcastle and North Tyneside (2, 4) and Lambeth (36).
When compared to ONS census 2011 published population statistics for the numbers of teenagers
aged 13-24 in each area (37), four of the five schemes gave costs between £0.33 and £0.68 per head
of teenage population per annum, with Lambeth having higher costs of £1.21 per head of teenage
population. In addition, as a sense check, bottom-up costings were developed with the help of
experts with experience of running schemes in both rural and urban areas. In these examples the
main costs were staff time to manage the scheme and train new providers and the costs of condoms
(and lube, although this was requested less often and is not offered in all c-card schemes). Additional
costs included the cost of the card itself, cost of an online registry (where used),
promotion/advertising costs (including web sites) and costs of travel and distribution. These costings
generated a possible range of costs of between £0.31 and £0.51 per head of teenage population. An
estimated cost of £0.48 per head of teenage population was chosen to be representative of the four
lower-cost published schemes (see Table 14), and a normal distribution with mean £0.48 and
standard error £0.07 applied in the PSA.
Table 14 C-card scheme costs
Parameter
Cost per head of
population (13-24)
31
Mean cost (SE)
£0.48 (£0.07)
Sources
(2-5)
2.4
Men who have sex with men (MSM) model
2.4.1 Economic model scope
This model was intended to capture the effects of implementing a CD scheme targeted at those with
a high risk of STIs, specifically MSM (ages 17-90). Such component schemes may be low-cost and aim
to increase availability of condoms to MSM of all ages, by providing access to free condom packs in a
variety of venues, especially those frequented by gay men, including pubs, clubs, shops and taxis. It
is hypothesised that these schemes would increase condom usage by reducing barriers to their
availability and purchase.
2.4.2 Effectiveness
The review identified only one single component scheme targeted at MSM (38), which was
considered a high quality before-and-after study. This study followed the London-based
‘RubberStuffers’ initiative in 1996, and surveyed MSM on their condom use before and after the
introduction of the scheme, which distributed up to 300,000 free condoms each year. Despite
improvements in possession of condoms, the study did not show a significant increase in the use of
condoms during anal intercourse. The review did include a US-based multi-component study
targeted at high-risk individuals including MSM (39), and this study did show increased condom
usage amongst those accessing the service (odds ratio = 1.37), suggesting that multi-component
schemes could be more effective, however, the review evidence was extremely limited, and the one
UK-based study is now quite old. Therefore, there was no evidence on which to base a model of a CD
scheme for MSM.
Due to a lack of effectiveness evidence, no base-case model was analysed. Instead a range of
effectiveness and cost options were assessed using a threshold analysis, to identify the combination
of costs and effectiveness required to make a scheme cost effective at £30,000 per QALY, or to make
a scheme dominant (QALY improving and cost-saving).
2.4.3 Model-specific parameters
2.4.3.1
Condom usage and failure
Data on condom use specific to MSM was available from the published 2008 UK Gay Men’s Sex
Survey (40). This online survey gave overall condom use at last anal sex within 6 months. Results
were broken down by region, and because the sample was heavily weighted towards London, the
32
results were weighted according to the overall English population to give an overall rate of condom
use of 52.7%. This was assumed to apply across all ages. Results of two more recent surveys (the
European MSM Internet Survey) were provided via personal communication from Ford Hickson1 and
confirmed that in both 2010 and 2014 condom use rates remained fairly stable in more recent years,
with a slight decrease in the numbers of men reporting always using a condom.
As discussed in the previous model, an overall failure rate of 3.62% (2.3% breakage and 1.3%
slippage) was used, based on Macaluso et al. (6).
2.4.3.2
Modelling sexually transmitted infections (STIs)
As well as data regarding the sexual activity and condom utilisation in a population, to determine the
number of STI cases in a population, the background population prevalence and transmission rate of
each STI are required.
2.4.3.3
Prevalence
For men who have sex with men the number of cases of Chlamydia, Gonorrhoea, HIV and Syphilis
diagnosed in 2014 was again taken from data published by Public Health England2. For Gonorrhoea,
HIV and Syphilis the number of cases diagnosed in 2014 among MSM was presented stratified by age
group. However for Chlamydia numbers were not presented for the whole of England although they
were presented for each Public Health England Centre (PHEC). The data for the London PHEC was
used to estimate the number of cases occurring in men who have sex with men in the whole of
England in 2014 on a pro-rata basis.
For MSM the total at risk population was again only presented stratified by age group. A value of
2.8%3 was used to estimate the number of men in the male population who had sex with other men.
Relevant data is presented in Table 15.
It was again assumed that the numbers given in the Public Health England data referred to the
number of cases reported in each age group and therefore the data was divided by the number of
1
Sigma Research, London School of Hygiene & Tropical Medicine
Public Health England. Table 2: STI diagnoses & rates by gender, sexual risk and age group, 2010 –
2014.London. Public Health England, Colindale
3. Erens B., et al. National Survey of Sexual Attitudes and Lifestyles II Reference tables and summary report.
2003. London. NATSAL, University College London.
2
33
years in each age group and further divided by the population in each at risk group to give the
percentage prevalence of Chlamydia, Gonorrhoea, HIV and Syphilis by age group for MSM. This
estimate of HIV prevalence was considered to the a lower estimate (being based on diagnoses), and
therefore two more HIV prevalence levels estimates were calculated; a central estimate using overall
prevalence of 5% (based on that estimated in the 2015 HIV in the UK Situation Report) (29), and a
higher estimate using overall prevalence of 9.1% (based on PHE’s update on HIV in MSM in London)
(41) – both assumed to spread across age groups according to the spread in incidence. These HIV
prevalence estimates are shown in Table 17.
Table 15: Prevalence of STI cases 2014 from Public Health England data for men who have sex with men
Age group
13 – 14
15 – 19
20 – 24
25 – 34
35 – 44
45 – 64
65+
Chlamydia
5
798
3,028
5,753
3,431
2,115
146
Men who have sex with men
Gonorrhoea
HIV
Syphilis
2
0
0
580
217
41
3,354
217
322
7,752
1,276
1,148
4,087
725
1,100
2,092
585
817
95
35
31
At risk population
17,610
46,779
51,222
102,827
99,653
187,939
117,205
Table 16: Mean prevalence values used in the model for Chlamydia, Gonorrhoea, HIV & Syphilis in the men who have sex
with men analysis
Age group
0 – 12
13 – 14
15 – 19
20 – 24
25 – 34
35 – 44
45 – 64
65+
34
Chlamydia
0.000%
0.026%
1.705%
5.911%
5.595%
3.443%
1.125%
0.124%
Gonorrhoea
0.000%
0.011%
1.240%
6.548%
7.539%
4.101%
1.113%
0.081%
HIV
0.000%
0.000%
0.464%
0.424%
1.241%
0.728%
0.311%
0.030%
Syphilis
0.000%
0.000%
0.088%
0.629%
1.116%
1.104%
0.435%
0.026%
Table 17 Lower, central and upper estimates of HIV prevalence in MSM
Age
group
13 – 14
15 – 19
20 – 24
25 – 34
35 – 44
45 – 64
65+
Lower
Central
Higher
0.000%
0.464%
0.424%
1.241%
0.728%
0.311%
0.030%
0.000%
4.732%
4.321%
12.657%
7.421%
3.175%
0.305%
0.000%
8.605%
7.859%
23.019%
13.496%
5.774%
0.554%
Pelvic Inflammatory Disease
Pelvic Inflammatory Disease is not relevant for the men who have sex with men
2.4.3.4
Transmission rate in the absence of condom use
Transmission rates for most STIs were assumed to be the same as in the general model. However,
for HIV evidence was available on transmission rates specifically for anal intercourse (42). See Table
18 for rates and sources.
Table 18 Transmission rate
STI
HIV (specific to anal intercourse)
Chlamydia
Gonorrhoea
Syphilis
35
Transmission rate (per 1 Source
sexual contact)
Men
Women
1.4%
Baggaley et al. (2010) (42)
45%
45%
Wang et al. (2000) (31)
53%
53%
Wang et al. (2000) (31)
61.8%
61.8%
Alexander et al. (1949) (32) (43)
(44)
2.5
Black African model
2.5.1 Economic model scope
This model was intended to capture the effects of implementing a scheme specifically targeted at
black Africans, who have higher prevalence of some STIs, including, importantly, HIV.
2.5.2 Effectiveness evidence
For this intervention no base-case model was analysed. Instead a range of effectiveness and cost
options were assessed using a threshold analysis, to identify the combination of costs and
effectiveness required to make a scheme cost effective at £30,000 per QALY, or to make a scheme
dominant (QALY improving and cost-saving).
2.5.3 Model-specific parameters
2.5.3.1
Condom usage and failure
Age-specific data for the proportion of people routinely using condoms were taken from an Office of
National Statistics (ONS) publication. (25) It was assumed that under 16s were the same as 16-19
year olds. These data are provided in Table 24.
Table 19 Condom use (percentage by age)
Age
16-19
20-24
25-29
30-34
35-39
40-44
45-49
Condom use (%)
54%
54%
41%
46%
27%
10%
13%
Source
Lader & Hopkins. ONS
Contraception and Sexual
health report, 2008/9 (25)
As discussed in the previous models, an overall failure rate of 3.62% (2.3% breakage and 1.3%
slippage) was used, based on Macaluso et al. (6).
2.5.4 Modelling sexually transmitted infections (STIs)
As well as data regarding the sexual activity and condom utilisation in a population, to determine the
number of STI cases in a population, the background population prevalence and transmission rate of
each STI are required.
2.5.4.1
Prevalence
Numbers of diagnoses of STIs in 2014 were available by ethnic group in PHE’s STI annual data tables
for chlamydia, gonorrhoea and syphilis (45). Numbers for the ‘black or black British’ group were
36
used. These were compared to 2011 census statistics for the number of people in the overall English
population who were black, African, Caribbean or black British to give rates of STI diagnosis. In the
absence of age-group-level data, the spread across age groups was assumed to be the same as for
STI diagnoses in the general population. Data are summarised in Table 20.
For HIV, three estimates of overall prevalence were used; 1.46% and 3.84% (men and women
respectively) were the lower level estimates, 1.79% and 4.55% were the central estimates and 2.33%
and 5.28% were the upper estimated (based on the 2015 HIV in the UK Situation Report (29) central
estimate and confidence interval, and consistent with the NICE HIV testing guideline1). In the
absence of age-specific prevalence estimates, rates were assumed to vary across age groups in the
same way as diagnoses in the general population. Data are summarised in Table 21.
Table 20 Mean estimated prevalence values used in the model for Chlamydia, Gonorrhoea & Syphilis in the black African
analysis
Age
group
13 – 14
15 – 19
20 – 24
25 – 34
35 – 44
45 – 64
65+
Chlamydia
Male
Female
0.015%
0.231%
1.491%
4.485%
3.044%
4.554%
1.192%
1.067%
0.346%
0.203%
0.096%
0.039%
0.010%
0.001%
Gonorrhoea
Male
Female
0.003%
0.028%
0.286%
0.469%
0.932%
0.454%
0.826%
0.157%
0.403%
0.041%
0.117%
0.010%
0.011%
0.000%
HIV
Male
0.231%
0.987%
0.901%
2.524%
2.026%
1.046%
0.228%
Female
0.641%
2.752%
2.457%
6.634%
5.297%
2.696%
0.492%
Syphilis
Male
Female
0.000%
0.000%
0.006%
0.003%
0.037%
0.005%
0.061%
0.005%
0.060%
0.002%
0.024%
0.001%
0.002%
0.000%
Table 21 Estimated lower, central and upper HIV prevalence for the black African analysis
Age
group
13 – 14
15 – 19
20 – 24
25 – 34
35 – 44
45 – 64
65+
1
Lower
Male
Female
0.231%
0.641%
0.987%
2.752%
0.901%
2.457%
2.524%
6.634%
2.026%
5.297%
1.046%
2.696%
0.228%
0.492%
Central
Male
Female
0.284%
0.759%
1.214%
3.261%
1.108%
2.911%
3.105%
7.861%
2.492%
6.277%
1.287%
3.195%
0.280%
0.583%
Upper
Male
Female
0.369%
0.881%
1.579%
3.784%
1.442%
3.378%
4.039%
9.122%
3.242%
7.283%
1.674%
3.707%
0.364%
0.676%
Available at: https://www.nice.org.uk/guidance/indevelopment/GID-PHG91/consultation/html-content
37
2.5.4.2
Transmission rate in the absence of condom use
Estimating the probability of transmission between an infected person and an uninfected person
from one unprotected sexual contact is difficult due to the ethical implications of this research.
Therefore the data for this important parameter are in some cases are old and limited. Rates and
sources are presented in Table 22.
Table 22 Transmission rate
STI
HIV
Chlamydia
Gonorrhoea
Syphilis
2.5.4.3
Transmission rate (per 1 sexual contact)
Men
Women
0.12%
0.39%
45%
45%
53%
53%
61.8%
61.8%
Source
Boily et al. (2009) (30)
Wang et al. (2000) (31)
Wang et al. (2000) (31)
Alexander et al. (1949)
(32) (43) (44)
Pelvic Inflammatory Disease
Pelvic Inflammatory Disease (PID) is an infection that often occurs during a case of chlamydia or
gonorrhoea in women. PID can have a significant impact on a women’s quality of life and in some
cases can be a serious condition requiring hospitalisation and surgery. Data were found providing
the proportion of chlamydia and gonorrhoea cases that result in PID.
Table 23 STIs that lead to PID
STI (leading to PID)
Chlamydia
Gonorrhoea
38
Proportion
16% (6 – 25%)
0.9% (0.1 - 1.8%)
Source
Price et al. (2013) (33)
Unpublished conference abstract (Australian data)
(46)
2.6
General population model
2.6.1 Economic model scope
This model was intended to capture the effects of implementing a population-wide CD scheme.
Schemes could include outlet schemes which provide condoms at a reduced price. An example of
such a scheme is the ‘Freedoms’ online shop operated by Central and North West London
Foundation Trust. This scheme provides a range of condoms and lubricants at a reduced price gained
through the purchasing power of the Trust, and is freely available to UK residents to use online for
home delivery. This scheme aims to improve condom use be reducing the cost barriers faced by
people purchasing small quantities of condoms locally.
2.6.2 Effectiveness evidence
The NICE evidence review identified only one study on the provision of discounted condoms, and
this study only reported outcomes for acquisition and not for condom use, therefore, no
effectiveness evidence is available. Therefore, for this intervention no base-case model was
analysed. Instead a range of effectiveness and cost options were assessed using a threshold analysis,
to identify the combination of costs and effectiveness required to make a scheme cost-effective at
£30,000 per QALY, or to make a scheme dominant (QALY improving and cost-saving).
2.6.3 Model-specific parameters
2.6.3.1
Condom usage and failure
Age-specific data for the proportion of people routinely using condoms were taken from an Office of
National Statistics (ONS) publication. (25) It was assumed that under 16s were the same as 16-19
year olds. These data are provided in Table 24.
Table 24 Condom use (percentage by age)
Age
16-19
20-24
25-29
30-34
35-39
40-44
45-49
Condom use (%)
54%
54%
41%
46%
27%
10%
13%
Source
Lader & Hopkins. ONS
Contraception and Sexual
health report, 2008/9 (25)
As discussed in the previous models, an overall failure rate of 3.62% (2.3% breakage and 1.3%
slippage) was used, based on Macaluso et al. (6).
39
2.6.4 Modelling sexually transmitted infections (STIs)
As well as data regarding the sexual activity and condom utilisation in a population, to determine the
number of STI cases in a population, the background population prevalence and transmission rate of
each STI are required.
2.6.4.1
Prevalence
See prevalence section of the young people model for prevalence of syphilis, gonorrhoea and
chlamydia. For HIV, three prevalence levels were estimated. The lower level prevalence was based
on diagnosis data (see section 2.3.4.1), whilst the central and upper estimates were based on overall
prevalences of 0.19% (based on the 2015 HIV in the UK Situation Report) (29) and 0.4% respectively
a value considered to represent “high prevalence” in the NICE HIV testing guideline1). In the absence
of age-specific rates, these were assumed to be spread across age groups in the same way as
incidence rates.
2.6.4.2
Transmission rate in the absence of condom use
Estimating the probability of transmission between an infected person and an uninfected person
from one unprotected sexual contact is difficult due to the ethical implications of this research.
Therefore the data for this important parameter are in some cases are old and limited. Rates and
sources are presented in Table 25.
Table 25 Transmission rate
STI
HIV
Chlamydia
Gonorrhoea
Syphilis
2.6.4.3
Transmission rate (per 1 sexual contact)
Men
Women
0.12%
0.39%
45%
45%
53%
53%
61.8%
61.8%
Source
Boily et al. (2009) (30)
Wang et al. (2000) (31)
Wang et al. (2000) (31)
Alexander et al. (1949)
(32) (43) (44)
Pelvic Inflammatory Disease
Pelvic Inflammatory Disease (PID) is an infection that often occurs during a case of chlamydia or
gonorrhoea in women. PID can have a significant impact on a women’s quality of life and in some
cases can be a serious condition requiring hospitalisation and surgery. Data were found providing
the proportion of chlamydia and gonorrhoea cases that result in PID.
1
Available at: https://www.nice.org.uk/guidance/indevelopment/GID-PHG91/consultation/html-content
40
Table 26 STIs that lead to PID
STI (leading to PID)
Chlamydia
Gonorrhoea
41
Proportion
16% (6 – 25%)
0.9% (0.1 - 1.8%)
Source
Price et al. (2013) (33)
Unpublished conference abstract (Australian data)
(46)
2.7
Summary of parameters
A full table of general model parameters is provided in Table 27
Table 27 General model parameters and distributions
Parameter
Value
SEXUAL PRACTICE – CONDOM USE (By age)
16-19
54%
20-24
54%
25-29
41%
30-34
46%
35-39
27%
40-44
10%
45-49
13%
SEXUAL PRACTICE – CONDOM USE (MSM)
All ages
52.7%
CONDOM
Breakage
3.6%
SEXUALLY ACTIVE – MEN
16-24
75.9%
25-34
90.1%
35-44
92.5%
45-54
86.4%
55-64
76.3%
6559.8%
SEXUALLY ACTIVE – WOMEN
16-24
77.0%
25-34
91.8%
35-44
90.8%
45-54
85.0%
55-64
63.7%
6542.1%
SEXUAL CONTACTS – MEN
16-24
5.10
25-34
5.40
35-44
4.10
45-54
4.10
55-64
3.20
652.30
SEXUAL CONTACT – WOMEN
16-24
5.80
25-34
4.90
35-44
4.00
45-54
3.50
55-64
2.50
651.40
PID
% chlamydia
16.0%
% gonorrhoea
0.9%
42
Distribution
(α,β) Rounded (unless specified)
Source
None
(25)
None
(40)
Beta (194, 9,704)
(6)
Beta (1,007, 320)
Beta (952, 105)
Beta (682, 55)
Beta (68, 11)
Beta (533, 166)
Beta (336, 226)
(14)
Beta (1,246, 372)
Beta (1,698, 152)
Beta (850, 86)
Beta (990, 175)
Beta (519, 296)
Beta (266, 365)
(14)
Gamma (0.50, 10.16)
Gamma (0.69, 7.82)
Gamma (0.91, 4.51)
Gamma (0.45, 9.08)
Gamma (0.51, 6.33)
Gamma (0.41, 5.63)
(14)
Gamma (0.77, 7.51)
Gamma (0.92, 5.31)
Gamma (0.76, 5.29)
Gamma (0.69, 5.04)
Gamma (0.54, 4.62)
Gamma (0.37, 3.78)
(14)
Beta (9, 47)
Beta (4, 469)
(33)
(46)
TRANSMISSION RATES
HIV – Men
0.120%
HIV - MSM
1.400%
HIV – Women
0.390%
Chlamydia
45.000%
Gonorrhoea
53.000%
Syphilis
61.818%
QALYs GAINED PER CASE AVERTED
Chlamydia
0.002
Gonorrhoea
0.004
HIV
4.450
Syphilis
0.003
PID
0.025
43
Beta (10, 8,175)
Beta (6, 394)
Beta (5, 1,324)
Beta (42, 52)
Beta (16, 14)
Beta (68, 42)
(30)
(42)
Normal (SD 0.0332)
Normal (SD 0.0001)
-
(16)
(16)
(15)
(17)
(18)
(31)
(31)
Condoms - health economics report
3
RESULTS
3.1
Young People (C-Card) scheme
The economic results for the C-Card scheme include 1) a base case scenario for the C-Card scheme
targeted at young people of age 13-24 years, presented in Table 28, with cost effectiveness plane
shown in Figure 3 2) a scenario analysis for a narrower age group population between 13-18 years
of age, presented in Table 29, with cost-effectiveness plane shown in Figure 4, 3) a scenario
including an increased effect on condom failure rate, Table 30, 4) a scenario with higher HIV
prevalence, Table 31 5) a two-way threshold analysis examining costs and effects of a multicomponent young people CD schemes, Table 32, and an analysis of potential pregnancy outcomes in
Table 33.
The baseline analysis targeted at the 13-24 years age group is in line with the generic C-Card
principles. For young people, the QoL effects of CD schemes are heavily influenced by the large
impact of a small number of HIV cases prevented. However, PID was also important. In terms of
costs, PID prevention was the most cost-saving effect of the scheme, followed by both HIV and
chlamydia. The model predicts that an intervention with effectiveness as per that demonstrated in
Larsson et al. (1) and with costs in the region of a typical C-card scheme would be expected to
increase QALYs and cost somewhat more than it saves, giving a cost of £17,411 per QALY gained.
However, it should be cautioned that the evidence for effectiveness in this broader age group is not
demonstrated, with the Larsson study being undertaken in the younger age group and specifically in
a school setting.
When a narrower age range for C-Card is considered, 13-18 years Table 29, the population is most
coherent with the study population of the Larsson study, from which the effectiveness is estimated
and the population is coherent with the estimated impact on unintended pregnancies. In this case
the reduced rates of sexual activity and underlying prevalence of STIs means that there is there is
less scope for savings, with cost savings only covering around half of what it costs to deliver,
meaning it would not be cost effective at the £30,000 level (ICER = £45,856).
The results of the scenario analysis where condom breakage was reduced (see Table 30) suggest that
although reducing condom breakage through education has some influence on the effectiveness and
cost effectiveness of the scheme (ICER reduces to £14,469), the impact of the kind of improvements
reported in the studies identified was not great.
In the scenario with higher prevalence of HIV, the increase in HIV cases averted by CD schemes
makes them cost-saving overall (£10m savings compared with £3.5m scheme costs across England in
the target population).
Table 33 shows the results of an exploratory analysis around pregnancy outcomes, including the
number of pregnancies and abortions avoided and the discounted cost savings associated with the
avoided pregnancies for a range of relative risks associated with the intervention (which correspond
to relative effects used within the two-way threshold analysis in Table 32). This includes Public
Sector costs associated with abortion, miscarriage, maternity, low birth weight babies and
government-funded Benefit payments. This analysis does not include the costs of the intervention or
the cost associated with STIs because these are included within the main STI analysis.
This analysis suggests that the costs saved from preventing or delaying a small number of
pregnancies in those aged under 18 creates significantly greater savings than those saved from STI
prevention. For example, at RR = 1.22 savings from STIs are estimated to be less than £764,000
whereas pregnancy-related savings are estimated at over £11m (including government benefit
savings), making the scheme cost-saving overall (saving over £10m). Even if benefit savings are
excluded from the pregnancy cost analysis, savings of around £318,000 are estimated, shifting the
ICER to around £12,000 for 13-24s.
45
Condoms - health economics report
Table 28: Probabilistic results obtained for multi-component (C-card) intervention in young people (13-24 years), effectiveness per Larsson et al. (20). Values (95% CI).
Cases per England
population in
target group
Conventional
Condom
distribution
Incremental cases
QALY gained
STI costs averted
Cost of CD scheme
Incremental cost
ICER
iNMB per person
Young people, multi-component intervention
Chlamydia
Gonorrhoea
HIV
Syphilis
PID
Total
64177 (30655,
85406)
59904 (26664,
84816)
-4272 (-10385, -361)
6122 (2727,
8174)
5744 (2290,
8129)
-378 (-931, -22)
27 (1, 84)
246 (53, 329)
22 (1, 68)
232 (51, 328)
-6 (-19, 0)
-14 (-38, 0)
7025 (1643,
13335)
6572 (1330,
13128)
-454 (-1432, -4)
77598 (37246,
103555)
72474 (31937,
102719)
-5123 (-12441, -439)
-8.43 (-20.5, -0.71)
-34 (-119, -2)
-0.09 (-0.24, 0)
-£568,408
-£2,943
-11.34 (-35.81, 0.1)
-£1,417,171
55 (14, 136)
-£520,867
-1.49 (-3.68, 0.09)
-£77,950
(-£1,266,135, £43,996)
(-£191,983, £4,622)
(-£1,979,739, £26,990)
(-£8,020, £088)
(-£4,475,253, £12,523)
(-£6,268,458, £597,461)
-£2,587,340
£3,544,962
£957,622
(-£2,723,496,
£2,947,501)
£17,411
£0.09
(-0.36, 0.83)
Figure 3: Cost-effectiveness plane for base-case analysis of a multi-component (C-card) intervention in young people (13-24 years), effectiveness per Larsson et al. (20).
47
Table 29: Probabilistic results obtained for multi-component (C-card) intervention in young people (13-18 years), effectiveness per Larsson et al. (20). Values (95% CI).
Cases per England
population of target
age
Conventional
Condom distribution
Incremental cases
QALY gained
STI costs averted
Cost of CD scheme
Incremental cost
ICER
iNMB per person
Young people, multi-component intervention
Chlamydia
Gonorrhoea
HIV
Syphilis
PID
Total
14453 (5686,
19936)
13302 (4669,
19781)
-1151 (-2857, 98)
-2.27 (-5.64, 0.19)
-£140,346
1121 (527,
1533)
1038 (415,
1527)
-83 (-212, -6)
7 (0, 21)
24 (10, 33)
17372 (6727, 24116)
5 (0, 16)
22 (8, 33)
-2 (-6, 0)
-2 (-5, 0)
1767 (344,
3433)
1632 (260,
3375)
-135 (-426, -1)
-0.33 (-0.84, 0.02)
-£17,170
-11 (-39, 0)
-0.01 (-0.03, 0)
17 (4, 44)
-£178,792
-£363
-3.38 (-10.64, 0.04)
-£422,276
(-£348,386, £11,911)
(-£43,771, £1,176)
(-£635,373, £7,706)
(-£955, -£014)
(-£1,329,660, £4,558)
15999 (5499, 23857)
-1373 (-3426, -117)
-£758,947
(-£1,850,255, -£166,793)
£1,538,499
£779,552
(-£311,756, £1,371,706)
£45,856
-£0.09
(-0.42, 0.42)
48
Figure 4: Cost-effectiveness plane for scenario analysis of a multi-component (C-card) intervention in young people (13-18 years), effectiveness per Larsson et al. (20).
49
Table 30 Probabilistic results obtained for the young people C-Card model where condom breakage is assumed to improve. Values (95% CI)
Cases per England
population in target
group
Conventional
Condom distribution
Incremental cases
QALY gained
STI costs averted
Larsson economic outcomes
Young people, multi-component intervention
Chlamydia
Gonorrhoea
HIV
Syphilis
PID
Total
62436 (28798,
84086)
57850 (24810,
83025)
-4586 (-10751, 557)
-9.05 (-21.23, 1.1)
-£559,107
5971 (2601,
8076)
5564 (2214,
8014)
-407 (-963, -38)
27 (1, 83)
241 (51, 327)
21 (1, 67)
226 (48, 326)
-5 (-19, 0)
-15 (-39, -1)
6839 (1459,
13123)
6352 (1184,
12809)
-487 (-1435, -8)
75514 (34522,
101875)
70013 (29689,
101013)
-5501 (-12926, -661)
-1.61 (-3.8, -0.15)
-34 (-117, -1)
-0.09 (-0.24, 0)
56 (14, 140)
-£83,902
-£559,691
-£3,150
-12.19 (-35.87, 0.19)
-£1,522,925
-£2,728,775
(-£198,568, £7,772)
(-£1,917,060, £22,424)
(-£8,254, -£149)
(-£4,482,192, £23,683)
(-£6,428,423, £638,582)
(-£1,310,817, £67,895)
Cost of CD scheme
Incremental cost
ICER
iNMB per person
£3,539,033
£810,258
(-£2,889,390,
£2,900,451)
£14,469
£0.12
(-0.34, 0.85)
50
Table 31: Probabilistic results obtained for multi-component (C-card) intervention in young people (13-24 years) with higher HIV prevalence (=0.19% across all age groups in population),
effectiveness per Larsson et al. (20). Values (95% CI).
Cases per England
population of target
age
Conventional
Condom distribution
Incremental cases
QALY gained
STI costs averted
Cost of CD scheme
Incremental cost
ICER
iNMB per person
Young people, multi-component intervention
Chlamydia
Gonorrhoea
HIV
Syphilis
PID
Total
64063 (30077,
85437)
59809 (25838,
84873)
-4254 (-10276, 381)
-8.4 (-20.29, 0.75)
-£518,701
6129 (2697,
8169)
5753 (2289,
8131)
-376 (-925, -24)
386 (19, 1217)
247 (53, 329)
77803 (36038, 104432)
309 (15, 974)
233 (51, 328)
-77 (-263, -3)
-14 (-38, 0)
6978 (1575,
13374)
6524 (1262,
13224)
-454 (-1450, -4)
-1.48 (-3.65, 0.1)
-£77,428
-475 (-1657, 18)
-£7,938,701
-0.09 (-0.24, 0)
-11.34 (-36.25, 0.09)
-£1,416,928
496 (40, 1688)
(-£1,252,842, £46,421)
(-£190,623, £5,010)
(-£27,113,377, £294,774)
(-£7,961, -£093)
(-£4,529,500, £11,011)
(-£29,089,132, -£1,719,544)
-£2,891
72629 (31047, 103456)
-5174 (-12500, -572)
-£9,954,650
£3,541,896
-£6,412,754
(-£25,547,236, £1,822,352)
Dominates
£2.86
(-0.06, 10.12)
51
CDS effectivenss (Relative Risk)
Table 32: Two way threshold analysis for the C-Card scheme in young people aged 13-24 years.
52
1.0001
1.020
1.040
1.060
1.080
1.100
1.120
1.140
1.160
1.180
1.200
1.220
1.240
1.260
1.280
1.300
1.400
1.500
CDS per person in target population costs
£
0.10 £
0.20 £
0.40
33,292,605
66,622,923
133,283,559
128,939
295,591
628,895
45,613
128,939
295,591
17,838
73,389
184,490
3,950
45,613
128,939
Dominates
28,948
95,609
Dominates
17,838
73,389
Dominates
9,902
57,517
Dominates
3,950
45,613
Dominates
Dominates
36,355
Dominates
Dominates
28,948
Dominates
Dominates
22,888
Dominates
Dominates
17,838
Dominates
Dominates
13,565
Dominates
Dominates
9,902
Dominates
Dominates
6,728
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
0.60
199,944,195
962,199
462,243
295,591
212,265
162,270
128,939
105,132
87,276
73,389
62,278
53,188
45,613
39,204
33,709
28,948
12,283
2,284
£
0.80
266,604,830
1,295,502
628,895
406,692
295,591
228,930
184,490
152,747
128,939
110,422
95,609
83,489
73,389
64,842
57,517
51,168
28,948
15,616
£
1.00
333,265,466
1,628,806
795,547
517,794
378,917
295,591
240,040
200,361
170,602
147,456
128,939
113,789
101,164
90,481
81,324
73,389
45,613
28,948
£
1.50
499,917,055
2,462,066
1,212,177
795,547
587,232
462,243
378,917
319,398
274,760
240,040
212,265
189,540
170,602
154,578
140,843
128,939
87,276
62,278
£
2.00
666,568,645
3,295,325
1,628,806
1,073,300
795,547
628,895
517,794
438,436
378,917
332,625
295,591
265,291
240,040
218,675
200,361
184,490
128,939
95,609
£
5.00
1,666,478,181
8,294,882
4,128,585
2,739,819
2,045,436
1,628,806
1,351,053
1,152,658
1,003,862
888,131
795,547
719,796
656,670
603,256
557,473
517,794
378,917
295,591
£
10.00
3,332,994,074
16,627,478
8,294,882
5,517,351
4,128,585
3,295,325
2,739,819
2,343,029
2,045,436
1,813,975
1,628,806
1,477,305
1,351,053
1,244,225
1,152,658
1,073,300
795,547
628,895
£
15.00
4,999,509,967
24,960,073
12,461,180
8,294,882
6,211,734
4,961,844
4,128,585
3,533,399
3,087,010
2,739,819
2,462,066
2,234,813
2,045,436
1,885,194
1,747,843
1,628,806
1,212,177
962,199
Table 33: Analysis of pregnancy outcomes in the C-Card scheme in English young people aged 13-18 years.
Incremental values
Relative risk of
condom use
Discounted cost savings associated with
Discounted cost savings associated
avoided pregnancies (excl. governmentwith avoided pregnancies (incl.
funded Benefits)
government-funded Benefits)*
1
0
0
£0
£0
1.02
10
7
£28,874
£1,020,652
1.04
21
14
£58,052
£2,051,477
1.06
31
21
£86,430
£3,060,737
1.08
41
28
£115,594
£4,089,983
1.1
51
36
£145,218
£5,134,459
1.12
62
43
£172,736
£6,114,544
1.14
72
50
£202,818
£7,158,562
1.16
82
57
£230,929
£8,172,669
1.18
92
64
£259,808
£9,200,348
1.2
103
71
£289,337
£10,231,813
1.22
113
78
£317,908
£11,228,242
1.24
123
85
£347,550
£12,262,177
1.26
133
92
£374,582
£13,233,075
1.28
144
100
£402,719
£14,249,880
1.3
154
107
£433,241
£15,299,207
1.4
206
142
£577,722
£20,429,179
1.5
257
178
£724,414
£25,525,243
*Note that the costs associated with government-funded Benefits are based on a model of the system from 2010 and uplifted.
53
No. of pregnancies
avoided
No. of abortions
avoided
Condoms - health economics report
3.2
MSM scheme
There was no base case for this scheme, since there was no evidence for effectiveness from the one
study in the review (38). Therefore, a range of scenarios were generated amongst the target
population (MSM aged 17-90), with varying effectiveness (relative risk of condom use) and cost per
person in the target population. The results are shown for lower, central and upper HIV prevalence
estimates (respectively) in Table 34, Table 35 and Table 36.
In this sub-group the increased QALYs and savings from preventing HIV, if a scheme is effective, are
much larger than the impacts from other STIs. The threshold analysis suggests that even schemes
with relatively high costs per person can be cost effective if a small improvement in condom use can
be achieved and demonstrated. For example, even at lower prevalence estimates, a scheme which
gave a relative risk of condom use of 1.04 compared with baseline would be cost effective even at
£10 per person in the target population (twenty times the cost suggested for a multi-component Ccard scheme).
Condoms - health economics report
CDS effectivenss (Relative Risk)
Table 34: Threshold analysis for MSM single component schemes: Lower HIV prevalence estimate()
1.0001
1.020
1.040
1.060
1.080
1.100
1.120
1.140
1.160
1.180
1.200
1.220
1.240
1.260
1.280
1.300
1.400
1.500
CDS per person in target population costs
£
0.10 £
0.20 £
0.40
164,881
343,058
695,875
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
0.60
1,067,622
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
0.80
1,426,396
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
1.00
1,797,072
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
1.50
2,668,920
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
2.00
3,548,726
1,176
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
5.00
9,029,812
27,961
5,609
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
10.00
18,169,377
74,106
27,770
13,041
5,637
1,131
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
15.00
27,114,332
117,437
50,936
28,403
16,704
9,990
5,702
2,504
104
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
CDS effectivenss (Relative Risk)
Table 35 Threshold analysis for MSM single component schemes: Central HIV prevalence estimate (5% overall)
56
1.0001
1.020
1.040
1.060
1.080
1.100
1.120
1.140
1.160
1.180
1.200
1.220
1.240
1.260
1.280
1.300
1.400
1.500
CDS per person in target population costs
£
0.10 £
0.20 £
0.40
1,284
18,841
54,429
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
0.60
90,376
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
0.80
124,823
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
1.00
161,886
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
1.50
248,028
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
2.00
342,455
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
5.00
880,937
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
10.00
1,741,197
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
15.00
2,655,130
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
CDS effectivenss (Relative Risk)
Table 36 Threshold analysis for MSM single component schemes: Upper HIV prevalence estimate (9.1% overall)
57
1.0001
1.020
1.040
1.060
1.080
1.100
1.120
1.140
1.160
1.180
1.200
1.220
1.240
1.260
1.280
1.300
1.400
1.500
CDS per person in target population costs
£
0.10 £
0.20 £
0.40
Dominates
3,063
22,375
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
0.60
42,625
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
0.80
61,614
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
1.00
81,013
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
1.50
129,427
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
2.00
179,681
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
5.00
478,179
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
10.00
973,233
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
15.00
1,444,593
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Condoms - health economics report
3.3
Black African scheme
There was no base case for this scheme, but a range of scenarios were generated amongst the target
population (English black African population aged 18-90), with varying effectiveness (relative risk of
condom use) and cost per person in the target population The results are shown for lower, central
and upper HIV prevalence estimates (respectively) in Table 37, Table 38 and Table 39.
In this population the increase in QALYs are mainly from HIV prevention, although in terms of cost
savings, HIV, PID and chlamydia prevention are all important. The threshold analysis suggests that
due to the higher prevalence of HIV in this sub-group, schemes could be cost-saving, if effective, at
relatively high cost. For example, even at the lower HIV prevalence level, a scheme with relative risk
of condom use of 1.04 would be cost effective even at £10 per person.
Condoms - health economics report
CDS effectivenss (Relative Risk)
Table 37 Threshold analysis for black African schemes: Lower HIV prevalence estimate
1.0001
1.020
1.040
1.060
1.080
1.100
1.120
1.140
1.160
1.180
1.200
1.220
1.240
1.260
1.280
1.300
1.400
1.500
CDS per person in target population costs
£
0.10 £
0.20 £
0.40
161,888
340,474
697,646
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
0.60
1,054,818
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
0.80
1,411,991
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
1.00
1,769,163
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
1.50
2,662,093
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
2.00
3,555,024
1,160
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
5.00
8,912,608
27,948
5,625
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
10.00
17,841,914
72,595
27,948
13,066
5,625
1,160
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
15.00
26,771,220
117,241
50,271
27,948
16,787
10,090
5,625
2,436
44
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
CDS effectivenss (Relative Risk)
Table 38 Threshold analysis for black African schemes: Central HIV prevalence estimate
60
1.0001
1.020
1.040
1.060
1.080
1.100
1.120
1.140
1.160
1.180
1.200
1.220
1.240
1.260
1.280
1.300
1.400
1.500
CDS per person in target population costs
£
0.10 £
0.20 £
0.40
133,342
283,376
583,442
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
0.60
883,509
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
0.80
1,183,575
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
1.00
1,483,642
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
1.50
2,233,808
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
2.00
2,983,974
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
5.00
7,484,971
20,817
2,063
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
10.00
14,986,634
58,326
20,817
8,315
2,063
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
15.00
22,488,296
95,834
39,572
20,817
11,440
5,814
2,063
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
CDS effectivenss (Relative Risk)
Table 39 Threshold analysis for black African schemes: Upper HIV prevalence estimate
61
1.0001
1.020
1.040
1.060
1.080
1.100
1.120
1.140
1.160
1.180
1.200
1.220
1.240
1.260
1.280
1.300
1.400
1.500
CDS per person in target population costs
£
0.10 £
0.20 £
0.40
110,702
238,089
492,863
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
0.60
747,637
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
0.80
1,002,411
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
1.00
1,257,186
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
1.50
1,894,121
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
2.00
2,531,056
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
5.00
6,352,668
15,162
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
10.00
12,722,021
47,008
15,162
4,546
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
15.00
19,091,374
78,855
31,085
15,162
7,200
2,423
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Condoms - health economics report
3.4
General population scheme
There was no base case for this scheme, since there was no outcome measure of condom use
effectiveness from the one study in the review (47). Therefore, a range of scenarios were generated
amongst the target population (English sexually active population aged 13-90), with varying
effectiveness (relative risk of condom use) and cost per person in the target population The results
are shown for lower, central and upper HIV prevalence estimates (respectively) in Table 40, Table 41
and Table 42.
In this broad population the increase in QALYs are mainly from HIV prevention, although in terms of
cost savings, HIV, PID and chlamydia prevention are all important. The threshold analysis suggests
that schemes need to balance cost and effectiveness to be cost effective. For example, at central HIV
prevalence estimates, a scheme with relative risk of condom use of 1.1 would have to cost around
£1 per person in the target population or less to be cost effective.
Condoms - health economics report
CDS effectivenss (Relative Risk)
Table 40: Threshold analysis for CD schemes targeted at the general population: Lower HIV prevalence ()
1.0001
1.020
1.040
1.060
1.080
1.100
1.120
1.140
1.160
1.180
1.200
1.220
1.240
1.260
1.280
1.300
1.400
1.500
CDS per person in target population costs
£
0.10 £
0.20 £
0.40
45,093,004
90,238,248
180,529,379
197,110
423,050
870,225
84,102
197,371
420,865
47,047
122,162
273,486
28,912
84,315
197,386
17,441
62,443
153,587
9,762
47,656
123,390
4,199
37,089
100,139
827
28,903
85,521
Dominates
22,900
72,649
Dominates
17,235
62,347
Dominates
13,274
54,799
Dominates
10,063
48,076
Dominates
7,009
41,514
Dominates
4,451
37,033
Dominates
2,333
31,976
Dominates
Dominates
17,295
Dominates
Dominates
8,200
£
0.60
270,743,869
1,329,326
649,413
421,762
311,623
241,227
195,365
165,504
140,912
120,802
106,951
95,736
84,521
75,958
68,500
62,342
39,913
26,529
£
0.80
362,149,160
1,765,912
870,874
574,711
423,399
329,952
273,906
226,404
198,046
172,939
151,624
134,364
123,741
110,447
101,152
91,467
62,410
44,860
£
1.00
447,053,070
2,226,064
1,104,938
725,647
533,802
425,239
348,235
294,151
252,327
221,141
197,328
179,286
159,065
144,242
132,572
122,861
84,594
62,664
£
1.50
675,823,225
3,373,326
1,667,907
1,110,163
811,254
647,470
537,798
455,341
396,028
349,708
309,046
279,860
252,665
232,385
212,874
197,127
141,398
108,465
£
2.00
897,766,652
4,473,812
2,234,500
1,479,861
1,094,492
870,494
720,084
613,240
536,258
466,266
425,100
383,846
348,633
314,356
294,825
274,054
198,962
151,959
£
5.00
2,270,674,332
11,307,582
5,620,262
3,735,135
2,779,229
2,236,085
1,844,342
1,569,707
1,375,460
1,220,012
1,092,768
990,487
908,226
833,412
774,247
723,949
535,690
423,252
£
10.00
4,480,802,760
22,463,198
11,094,479
7,459,136
5,574,725
4,473,834
3,716,577
3,177,021
2,773,066
2,493,551
2,231,365
2,021,898
1,836,383
1,692,525
1,570,416
1,465,745
1,102,418
872,662
£
15.00
6,755,301,498
33,649,274
16,830,213
11,244,167
8,351,710
6,736,111
5,604,552
4,810,891
4,141,676
3,705,671
3,348,660
3,020,338
2,767,651
2,564,649
2,368,399
2,229,114
1,663,972
1,324,922
CDS effectivenss (Relative Risk)
Table 41 Threshold analysis for CD schemes targeted at the general population: Central HIV prevalence (0.19% overall)
64
1.0001
1.020
1.040
1.060
1.080
1.100
1.120
1.140
1.160
1.180
1.200
1.220
1.240
1.260
1.280
1.300
1.400
1.500
CDS per person in target population costs
£
0.10 £
0.20 £
0.40
3,715,468
7,409,609
14,754,927
931
19,543
56,866
Dominates
1,066
19,257
Dominates
Dominates
7,204
Dominates
Dominates
1,067
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
0.60
22,129,498
93,915
38,681
19,062
10,200
4,752
1,009
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
0.80
29,617,109
132,395
57,031
32,219
19,223
12,237
7,043
3,616
1,201
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
1.00
37,246,043
166,321
75,083
44,080
28,210
19,519
13,249
8,802
5,862
3,262
1,005
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
1.50
56,298,477
261,764
122,006
76,309
51,652
38,505
28,031
22,290
17,135
13,467
10,087
7,970
5,522
3,607
2,142
1,278
Dominates
Dominates
£
2.00
73,638,607
347,095
170,174
105,849
75,102
56,642
44,544
34,881
29,165
23,861
19,754
16,015
13,203
11,095
8,976
7,661
843
Dominates
£
5.00
185,243,726
910,531
448,922
291,688
215,078
167,762
135,570
114,464
98,767
86,124
75,553
66,105
59,721
53,426
49,689
44,267
29,256
19,617
£
10.00
371,713,406
1,829,294
909,269
595,369
450,422
354,856
293,468
246,281
215,494
192,666
169,814
151,689
135,765
125,625
113,749
105,644
75,007
56,904
£
15.00
561,551,635
2,754,098
1,359,076
915,564
677,671
538,632
441,570
378,549
330,801
292,149
262,331
237,371
217,405
198,736
179,964
166,715
122,172
94,926
CDS effectivenss (Relative Risk)
Table 42 Threshold analysis for CD schemes targeted at the general population: Upper HIV prevalence (0.4% overall)
65
1.0001
1.020
1.040
1.060
1.080
1.100
1.120
1.140
1.160
1.180
1.200
1.220
1.240
1.260
1.280
1.300
1.400
1.500
CDS per person in target population costs
£
0.10 £
0.20 £
0.40
1,788,497
3,558,681
7,025,437
Dominates
814
18,608
Dominates
Dominates
658
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
0.60
10,488,671
36,472
9,237
820
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
0.80
14,264,087
54,451
18,587
6,566
766
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
1.00
17,839,848
71,018
27,238
12,334
4,811
837
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
1.50
26,523,398
114,772
50,713
27,549
16,161
9,495
5,031
1,835
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
2.00
35,162,493
162,334
72,202
42,058
27,165
18,243
12,673
8,337
5,074
2,541
769
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
Dominates
£
5.00
87,815,743
421,063
206,271
131,150
93,847
70,678
57,363
46,197
38,730
31,969
27,037
23,542
19,882
17,415
14,425
12,684
5,088
580
£
10.00
176,707,108
877,372
424,261
278,252
202,410
157,821
131,108
109,139
93,926
81,172
71,368
62,630
57,151
50,644
46,091
41,794
27,595
18,389
£
15.00
265,499,729
1,317,323
649,786
423,941
310,808
247,455
205,127
172,412
150,613
130,388
115,259
104,976
93,599
84,959
77,372
70,651
49,478
35,875
Condoms - health economics report
4
DISCUSSION AND CONCLUSION
4.1
Discussion of results
Overall the economic evidence base for CD schemes identified in the review is poor, with a heavy
reliance on studies undertaken in the USA with potentially little applicability to the UK population.
The results of the four models presented here suggest that HIV prevention is the primary driver of
health benefits and cost savings from CD schemes, with important secondary effects arising through
PID prevention in schemes which target both sexes, and chlamydia prevention. The analysis suggests
that schemes that target higher risk populations are more likely to be cost effective at higher cost
and lower effectiveness. For the general population, a CD scheme with a relative risk of 1.10 would
need to cost less than 20p per person to be cost effective in a low prevalence HIV population, but
would be cost effective at £2 per person in a high prevalence HIV population. CD schemes for MSM
and black African populations are more cost effective because of the higher prevalence of HIV, and
in MSM the higher risk of transmission through anal intercourse. In the case of teenagers the higher
prevalence of chlamydia in this subgroup increases the cost effectiveness of CD schemes.
For young people, restricting the analysis to STI prevention outcomes only, and assuming
effectiveness in line with the best quality available study, a typical C-Card scheme would be cost
effective at the £30,000 per QALY level but not cost saving. However, taking into account the cost
savings associated with preventing or delaying pregnancies in teenagers these schemes are expected
to be highly cost effective (and cost-saving if government benefits are included in the analysis). CCard schemes are also intended to have a range of other STI-related benefits which have not been
explicitly considered here, such as linking young people up with sexual health services to improve
screening rates and improve prevention and treatment later in life.
For MSM, although the evidence was that single-component schemes are not effective in changing
condom-use behaviour, the threshold analysis suggested that schemes targeted at this high-risk
group could be cost effective even if considerably more expensive than the C-Card multi-component
scheme. NICE’s evidence review did include one multi-component study targeted at high risk of HIV
individuals (39) (not just MSM) which was effective at increasing reported condom use. Therefore, in
this group it is possible that more effective, if costlier, multi-component schemes could be the most
cost effective.
For black Africans, threshold analysis suggested that CD schemes targeting this high risk population
could be cost effective even at relatively high costs.
For the general population, threshold analysis suggested that CD schemes could be cost effective if
they could be delivered at a relatively low cost, and have some effect. Outlet schemes that provide
condoms at a reduced cost could be an example of such a scheme.
4.2
Limitations and further research
The economic impact of CD schemes was based on very limited effectiveness evidence, with a high
degree of variability in CD scheme designs, costs and effects being suggested.
There were a number of limitations to the model and results presented here which should be borne
in mind. In particular, because of the sensitive nature of the topic and the challenges around
conducting research in this area, obtaining robust, recent evidence for many of the model
parameters was difficult. For example, condom usage data were the same for young people as the
general population, but could be quite different in reality.
For the C-card model, evidence was taken from a Swedish, school-based study. Although the study
was of high quality, the intervention is not an exact match for the C-Card scheme, which is delivered
quite differently, through a range of sites, and with a broader group of young people. This study also
reported outcomes for reported ‘ever condom used’ versus ‘never used’ which could be overestimating the level of effectiveness for regular condom use. In addition, C-Card costs were shown to
be quite variable between areas, reflecting the differences in the way these schemes are delivered.
The variability around typical scheme costs was reflected in the model through PSA, however, there
do appear to be a number of schemes running at much higher costs, which could be much less cost
effective. It was not clear why these schemes are more expensive, although all schemes incur
different types of costs – for example, the bottom up costs we used for validation did not include
any payments to outlets (e.g. pharmacies) for being part of the scheme. However, we did come
across evidence of schemes that are paying retainers and fees to pharmacies in order to include
them in the scheme.
Although the costs and health impacts of PID were included, it is known that repeated episodes of
PID can increase the risk of infertility in women. The effects, either in terms of costs or quality of life
were not considered in the model, but should be considered as wider benefits of PID prevention.
For the MSM model, HIV costs dominated the results. However, it should be noted that the costs
were only available use present costs discounted at 3.0% (compared to NICE’s recommendation of
3.5%) which could lead to slight overestimates in the HIV cost estimates.
67
Finally, the model itself simplifies the process of disease transmission. It assumes that every act of
sexual intercourse is exposing a person to an STI, which may lead to an overestimate in the total
number of STI cases, since in individual in a monogamous relationship with someone who does not
have an STI is not at risk of transmission, even if they have unprotected sex. The model further
assumes that there is no relationship between sexual activity level and STI prevalence level for a
given person – the impact of this on STI transmission is unclear.
Using a static model assumes that over the model time horizon, at each act of sexual intercourse,
there is a constant proportion of the population with each STI. In reality, for STIs from which people
recover, if the recovery period is shorter (longer) than the average duration between acts of sexual
intercourse, it is possible that the model overestimates (underestimates) the increase in STI
prevalence. For STIs from which people do not recover (such as HIV), there will be an increased
proportion of the population with the STI after each act of sexual intercourse. The model therefore
likely underestimates the increase in HIV prevalence. This underestimate is more pronounced
where sexual activity levels and HIV prevalence is higher.
In order to understand the economics of CD schemes in the UK it is imperative to have evaluations
that demonstrate the impact of these schemes on behaviour change and condom usage in UKrelevant populations.
68
5
REFERENCES
1.
Larsson M, Eurenius K, Westerling R, Tyden T. Evaluation of a sexual education intervention
among Swedish high school students. Scand J Public Health. 2006;34(2):124-31.
2.
Jablonskas S. Nottinghamshire Condom Card (“C Card”) Distribution Scheme Department of
Health, 2010.
3.
Lincolnshire County Council Children’s Services in Partnership with NHS Lincolnshire. C Card
Condom Scheme – Evaluation. 2010.
4.
Newcastle & North Tyneside Health Promotion Department. C-Card Condom Distribution
Scheme Newcastle and North Tyneside 2000.
5.
Varley E. REPORT OF THE FINDINGS OF THE REVIEW OF DERBYSHIRE SEXUAL HEALTH
SERVICES (Health and Communities) DERBYSHIRE COUNTY COUNCIL 2014.
6.
Macaluso M, Kelaghan J, Artz L, Austin H, Fleenor M, Hook EW, 3rd, et al. Mechanical failure
of the latex condom in a cohort of women at high STD risk. Sex Transm Dis. 1999;26(8):450-8.
7.
Ellis S, Ciullum A, Carmona C, Kavanagh J, Murray A, Shaw T. Sexually transmitted infections:
condom distribution schemes. NICE, 2016 Contract No.: PHAC A 21.5 evidence review.
8.
NICE. Developing NICE guidelines: the manual. 2014.
9.
Pinkerton S, Abramson P. Evaluating the Risks A Bernoulli Process Model of HIV Infection and
Risk Reduction. Evaluation Review. 1993.
10.
Sukthankar A. Syphilis. Medicine. 2010;38(5):263-6.
11.
NICE. Contraceptive services for under 25s. 2014.
12.
Nherera L, Jacklin P. A model to assess the cost-effectiveness of Sex and Relationship
Education (SRE) developed for NICE public health guidance on personal, social, health and economic
(PSHE) education. National Collaborating Centre for Women’s and Children’s Health, 2009.
13.
Fenwick E, Claxton K, Sculpher M. Representing uncertainty: the role of cost-effectiveness
acceptability curves. Health economics. 2001;10(8):779-87.
14.
Mercer C, Tanton C, Prah P, Erens B. Changes in sexual attitudes and lifestyles in Britain
through the life course and over time: findings from the National Surveys of Sexual Attitudes and
Lifestyles. The Lancet. 2013.
15.
Farnham P, Gopalappa C. Updates of lifetime costs of care and quality-of-life estimates for
HIV-infected persons in the United States: late versus early diagnosis and entry into care. JAIDS
Journal of Acquired Immune Deficiency Syndromes. 2013.
16.
Turner K, Round J, Horner P. An early evaluation of clinical and economic costs and benefits
of implementing point of care NAAT tests for Chlamydia trachomatis and Neisseria gonorrhoea.
Sexually transmitted infections. 2013.
17.
Tuite AR, Burchell AN, Fisman DN. Cost-effectiveness of enhanced syphilis screening among
HIV-positive men who have sex with men: a microsimulation model. PloS one. 2014;9:e101240.
18.
Looker KJ, Wallace LA, Turner KME. Impact and cost-effectiveness of chlamydia testing in
Scotland: a mathematical modelling study. Theoretical biology & medical modelling. 2015;12:2.
19.
Agency HP, editor HIV in the United Kingdom: 2011 Report2011; London.
20.
The UK Faculty of Public Health. FPH submission to the House of Lords Select Committee on
HIV and AIDS in the UK – Call for evidence. 2011.
21.
Nakagawa F, Miners A, Smith CJ, Simmons R, Lodwick RK, Cambiano V, et al. Projected
Lifetime Healthcare Costs Associated with HIV Infection. PloS one. 2015;10:e0125018.
22.
Curtis L, Burns A. Unit Costs of Health and Social Care 2015 Personal Social Services Research
Unit, 2015.
23.
Guttmacher S, Lieberman L, Ward D, Freudenberg N, Radosh A, Des Jarlais D. Condom
availability in New York City public high schools: relationships to condom use and sexual behavior.
Am J Public Health. 1997;87(9):1427-33.
69
24.
Furstenberg FF, Jr., Geitz LM, Teitler JO, Weiss CC. Does condom availability make a
difference? An evaluation of Philadelphia's health resource centers. Fam Plann Perspect.
1997;29(3):123-7.
25.
Lader D, Hopkins G. Contraception and sexual health: a report on research using the
National Statistics Omnibus Survey produced on behalf of the NHS information centre for. Office for
National Statistics, Omnibus Survey Report. 2008.
26.
Messiah A, Dart T, Spencer BE, Warszawski J. Condom breakage and slippage during
heterosexual intercourse: a French national survey. French National Survey on Sexual Behavior
Group (ACSF). Am J Public Health. 1997;87(3):421-4.
27.
Frezieres RG, Walsh TL, Nelson AL, Clark VA, Coulson AH. Evaluation of the efficacy of a
polyurethane condom: results from a randomized, controlled clinical trial. Fam Plann Perspect.
1999;31(2):81-7.
28.
Vessey M, Lawless M, Yeates D. Efficacy of different contraceptive methods. Lancet.
1982;1(8276):841-2.
29.
Skingsley A, Yin Z, Kirwan P, Croxford S, Chau C, Conti S, et al., editors. HIV in the UK –
Situation Report 2015: data to end 2014. November 20152015; London.
30.
Boily M-C, Baggaley RF, Wang L, Masse B, White RG, Hayes RJ, et al. Heterosexual risk of HIV1 infection per sexual act: systematic review and meta-analysis of observational studies. The Lancet
Infectious Diseases. 2009;9:118-29.
31.
Wang L, Davis M. Economic evaluation of Safer Choices: a school-based human
immunodeficiency virus, other sexually transmitted diseases, and pregnancy prevention. JAMA
Pediatrics. 2000.
32.
Alexander LJ, Schoch AG. Prevention of syphilis: Penicillin Calcium in Oil and White Wax,
USP, Bismuth Ethylcamphorate and Oxophenarsine Hydrochloride in Treatment, During. Archives of
Dermatology. 1949.
33.
Price MJ, Ades AE, De Angelis D, Welton NJ, Macleod J, Soldan K, et al. Risk of pelvic
inflammatory disease following Chlamydia trachomatis infection: analysis of prospective studies with
a multistate model. American journal of epidemiology. 2013;178:484-92.
34.
Abortion statistics, England and Wales [Internet]. 2014. Available from:
https://www.gov.uk/government/statistical-data-sets/abortion-statistics-england-and-wales-2014.
35.
Consumer Price inflation time series dataset [Internet]. 2016 [cited 17.04.2016]. Available
from: https://www.ons.gov.uk/economy/inflationandpriceindices/datasets/consumerpriceindices.
36.
Lambeth Southwark and Lewisham. Sexual Health Strategy 2014-2017. 2014.
37.
2011 Census: Population estimates by single year of age and sex for local authorities in the
United
Kingdom
[Internet].
2011
[cited
17.04.2016].
Available
from:
https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populationesti
mates/datasets/2011censuspopulationestimatesbysingleyearofageandsexforlocalauthoritiesintheuni
tedkingdom.
38.
Weatherburn P. Distribution of free condom packs via gay commercial venue in London: An
evaluation of 'Rubberstuffers' pilot scheme. Sigma Research, 1998.
39.
Wendell DA, Cohen DA, LeSage D, Farley TA. Street outreach for HIV prevention:
effectiveness of a state-wide programme. Int J STD AIDS. 2003;14(5):334-40.
40.
Sigma Research. The UK Gay Men’s Sex Survey (part of EMIS - the European MSM Internet
Sex survey) London School of Hygiene & Tropical Medicine., 2010 March 2011. Report No.
41.
Public Health England. Inequalities in sexual health: Update on HIV and STIs in men who
have sex with men in London. 2016.
42.
Baggaley RF, White RG, Boily MC. HIV transmission risk through anal intercourse: systematic
review, meta-analysis and implications for HIV prevention. Int J Epidemiol. 2010;39(4):1048-63.
43.
Gray TG, Powles E. Understanding and managing syphilis. InnovAiT: Education and
inspiration for general practice. 2013;6(12):781-9.
70
44.
Singh AE, Romanowski B. Syphilis: Review with Emphasis on Clinical, Epidemiologic, and
Some Biologic Features. Clinical Microbiology Reviews. 1999;12(2):187-209.
45.
Public Health England. Sexually transmitted infections (STIs): annual data tables. In: Public
Health England, editor. 2014.
46.
Goller J, Fairley C, Bradshaw C, De Livera A, Chen M, Guy R, et al. RISK OF PELVIC
INFLAMMATORY DISEASE FROM
CHLAMYDIA AND GONORRHOEA AMONG AUSTRALIAN
SEXUAL HEALTH CLINIC ATTENDEES ISSTDR Conference2015.
47.
Dahl DW, Gorn GJ, Weinberg CB. Encouraging use of coupons to stimulate condom
purchase. Am J Public Health. 1999;89(12):1866-9.
71
Appendix A. C-card costings
To cost the C-card scheme a rapid search identified a small number of published documents from local schemes in England and Wales, of which 5 provided
overall costs of their schemes: Derbyshire (5), Lincolnshire (3), Nottinghamshire (2), Newcastle and North Tyneside (2, 4) and Lambeth (36).
Four of the documents gave annual costs, either for a single year (Derbyshire, Newcastle and Lambeth) or for multiple years (6 years, Nottinghamshire). For
Lincolnshire, the cost per annum of new registrants was provided, and combined with the stated number of new registrants, annual total scheme costs
were calculated. Stated annual costs were inflated to 2015 prices using the CPI inflation index, and converted to cost per person in the target population
(young people aged 13-24) using local authority, age-specific Census data from 2011.
When compared to ONS census 2011 published population statistics for the numbers of teenagers aged 13-24 in each area (37), four of the five schemes
gave costs between £0.33 and £0.68 per head of teenage population per annum, with Lambeth having higher costs of £1.21 per head of teenage
population.
Target Population
(people aged 13 to
24, Census 2011)
Area
Lincolnshire
Nottinghamshire
Derbyshire
Newcastle
&
North Tyneside
Cost per new
registrant
Total scheme
cost
Cost
year
2015 Total
scheme
cost
Cost per
person in
target pop
102,260
111,444
106,846
2,571*
3,277*
N/A
£15.08*
N/A
N/A
£38,766
£33,413*
£71,000*
2010
2010
2013
£43,843
£37,789
£72,609
£0.43
£0.34
£0.68
91,400
4,344*
N/A
£30,000*
2000
£41,433
£0.45
N/A
N/A
£56,000*
2014
£56,170
£1.20
Lambeth
46,622
All costs are per annum
*published value (others are calculated)
72
New
registrations
per annum
Source
Lincolnshire
County
Council
Children’s Services in Partnership
with NHS Lincolnshire, 2010
Jablonskas S., 2010
Varley E., 2014
Newcastle & North Tyneside Health
Promotion Department, 2000.
Lambeth Southwark and Lewisham,
2014.